01-projects/self-review

review log

2026-04-13·project·status: active

Vault Self-Review Log

Review 1 — 2026-04-13

Reviewer: Ray (AI COO) Scope: 5 entries from 06-reference/ filed on 2026-04-13 Max score: 13

Scored results

# File FM (2) Why (3) Map (3) Links (2) Bias (1) Walls (1) Concise (1) Total Grade
1 2026-04-13-mg-harness-review-cc-wrapped.md 2 3 3 2 1 1 1 13 A
2 [[2026-04-13-stratechery-mythos-muse-compute]] 2 0 3 2 1 1 1 10 B
3 [[book-solve-everything-master-synthesis-2026-04-13 2026-04-13-solve-everything-master-synthesis]] 2 0 3 2 1 1 1 10
4 [[2026-04-13-langchain-evals-deep-agents]] 2 3 3 2 1 1 1 13 A
5 [[2026-04-13-joe-reis-ai-hard-parts]] 2 0 3 2 1 1 1 10 B

Grade distribution

Grade Count Entries
A (12-13) 2 mg-harness-review, langchain-evals
B (10-11) 3 stratechery-mythos, solve-everything-master, joe-reis
C (8-9) 0 —
D (<8) 0 —

Summary statistics

Top entries

  1. mg-harness-review-cc-wrapped.md (13/13) — Complete on every criterion. Has explicit why-in-vault, detailed RDCO mapping with actionable takeaways, and verified cross-links.
  2. langchain-evals-deep-agents.md (13/13) — Strong why-in-vault tied to /improve skill, explicit bias flags section beyond what's required, clean RDCO mapping with four connection points.

Entries needing attention

None scored C or below. Three entries scored B due to the same missing element (see systemic patterns).

Systemic patterns

1. Missing "Why this is in the vault" section (3 of 5 entries) This is the dominant quality gap. Three entries — stratechery-mythos, solve-everything-master, joe-reis — lack an explicit why-in-vault section. All three have strong RDCO mapping, so the justification is implicitly present, but the criterion requires a dedicated section. The why-in-vault section serves a different purpose than the mapping: it answers "why did we file this at all" before diving into "how does it connect." Without it, a future reader has to infer relevance from the mapping section.

Action: Update the filing SOP (process-newsletter skill and any intake templates) to enforce a ## Why this is in the vault section as a required heading. One to three sentences, before any content summary.

2. Cross-link path inconsistency (minor) One cross-link in langchain-evals ([[cross-check-agent-architecture]]) uses a bare filename without the date prefix or subdirectory path. The actual file is [[2026-04-12-cross-check-agent-architecture]]. Obsidian's shortest-path resolution may handle this, but it's fragile — if another file with a similar name appears, the link breaks.

Action: Prefer full date-prefixed filenames in wikilinks. Not urgent but worth standardizing.

3. All entries pass on conciseness, bias flagging, and no-copy-paste No drift detected on these criteria. The filing process is producing clean, original synthesis at reasonable length. The two newsletter entries with sponsorship both have sponsored fields correctly populated, and joe-reis even includes a dedicated bias-notes section — exemplary.

4. Mapping quality is uniformly high Every entry scored 3/3 on the mapping criterion. The RDCO connections are specific, reference other vault entries by name, and propose concrete actions or position implications. This is the strongest dimension across the batch.

Process recommendations

  1. Add ## Why this is in the vault as a mandatory heading in filing templates — this is the only criterion dragging scores from A to B.
  2. Consider whether the solve-everything-master-synthesis format (no why-in-vault, but extensive positional mapping) deserves its own template — book-synthesis entries may warrant different structure than article-processing entries.
  3. No entries need remediation. The three B-scored entries could be upgraded to A by adding a two-sentence why-in-vault section, but that's a process fix going forward, not a backfill priority.

Review 2 — 2026-04-19

Reviewer: Ray (AI COO) Scope: 30 entries from 06-reference/ modified in last 7d (2026-04-12 → 2026-04-19), --fix mode Max score: 13

Scored results

# File FM (2) Why (3) Map (3) Links (2) Bias (1) Walls (1) Concise (1) Total Grade
1 [[synthesis-harness-thesis-dissent-2026-04-12 2026-04-12-harness-thesis-dissent]] 2 3 3 2 1 1 1 13
2 [[2026-04-19-kingsbury-future-of-everything-is-lies]] 2 3 3 2 1 1 1 13 A
3 [[2026-04-19-garry-tan-build-the-car-jepsen-response]] 2 3 3 2 1 1 1 13 A
4 [[2026-04-19-lia-dibello-academic-papers]] 2 3 3 2 1 1 1 13 A
5 [[2026-04-19-newsletter-platform-sanity-check-v3]] 2 3 3 2 1 1 1 13 A
6 [[2026-04-19-mac-vs-published-data-quality-frameworks]] 2 3 3 2 1 1 1 13 A
7 [[backfill-wrapup-commoncog-startherewrap-and-triad-2026-04-19 2026-04-19-commoncog-startherewrap-and-triad]] 2 3 3 2 1 1 1 13
8 [[2026-04-19-commoncog-framework-mental-models-to-practice]] 2 3 3 2 1 1 1 13 A
9 [[2026-04-19-commoncog-what-the-ceo-wants-you-to-know]] 2 3 3 2 1 1 1 13 A
10 [[2026-04-19-commoncog-user-review-procrastination-equation]] 2 3 3 2† 1 1 1 13 A (post-fix)
11 [[2026-04-19-commoncog-update-perceptual-exposure-learning]] 2 3 3 2 1 1 1 13 A
12 [[2026-04-19-commoncog-ultimate-guide-reading-book-a-week]] 2 3 3 0 1 1 1 11 B
13 [[2026-04-19-commoncog-tacit-skill-in-wicked-domains]] 2 3 3 2 1 1 1 13 A
14 [[2026-04-19-commoncog-seth-godin-the-dip]] 2 3 3 2† 1 1 1 13 A (post-fix)
15 [[2026-04-19-commoncog-reading-quickly-reading-lots]] 2 3 3 2† 1 1 1 13 A (post-fix)
16 [[2026-04-19-commoncog-reading-program-b2b-sales]] 2 3 3 2 1 1 1 13 A
17 [[2026-04-19-commoncog-product-validation-taste]] 2 3 3 2 1 1 1 13 A
18 [[2026-04-19-commoncog-product-development-iterated-taste]] 2 3 3 2 1 1 1 13 A
19 [[2026-04-19-commoncog-playlist-of-awesome-perceptual-exposure]] 2 3 3 2 1 1 1 13 A
20 [[2026-04-19-commoncog-personal-brand-as-moat-soft-landing]] 2 3 3 2 1 1 1 13 A
21 [[2026-04-19-commoncog-obviously-awesome]] 2 3 3 2 1 1 1 13 A
22 [[2026-04-19-commoncog-nuanced-take-preventing-burnout]] 2 3 3 2† 1 1 1 13 A (post-fix)
23 [[2026-04-19-commoncog-map-of-expertise-research]] 2 3 3 2 1 1 1 13 A
24 [[2026-04-19-commoncog-loose-feedback-loop]] 2 3 3 2 1 1 1 13 A
25 [[2026-04-19-commoncog-lia-dibello-business-expertise]] 2 3 3 2 1 1 1 13 A
26 [[2026-04-19-commoncog-land-and-expand-strategy-reading]] 2 3 3 2† 1 1 1 13 A (post-fix)
27 [[2026-04-19-commoncog-in-defence-of-reading-goals]] 2 3 3 2† 1 1 1 13 A (post-fix)
28 [[2026-04-19-commoncog-hold-lessons-of-history-loosely]] 2 3 3 2 1 1 1 13 A
29 [[2026-04-19-commoncog-gap-reputation-personal-brand]] 2 3 3 2 1 1 1 13 A
30 [[2026-04-19-commoncog-follow-your-nose]] 2 3 3 2† 1 1 1 13 A (post-fix)

† = Links criterion passed only after auto-fix added cross-links to same-cohort siblings.

Grade distribution (post-fix)

Grade Count %
A (12-13) 29 96.7%
B (10-11) 1 3.3%
C (8-9) 0 0%
D (<8) 0 0%

Summary statistics

Top entries (template-grade)

Worst entry (post-fix)

[[2026-04-19-commoncog-ultimate-guide-reading-book-a-week]] — 11/13. Single Apr-15 cross-link, was not auto-patched because it serves as the canonical anchor that the other 6 reading-category entries now cross-link TO. Consider hand-adding 1-2 links to other Apr-19 reading entries on next pass. Not a substantive issue.

Systemic patterns

1. Cohort-backfill template repetition (the dominant pattern this batch). All 23 Apr-19 Commoncog entries share start_here_category-keyed Why and Mapping sections — the text is RDCO-specific (passes the rubric) but identical across siblings. A future reader sees the same paragraph 7+ times across the Tacit-Knowledge category. The /tmp/cc_process_article.py + /tmp/cc_claims.py pipeline that produced the cohort lives outside ~/.claude/skills/ and was not subject to per-article specificity discipline.

Action: Filed Notion task "/improve: cohort backfill skill should individualize why-in-vault and mapping per article" (page id 347f7d49-36d1-81eb-b0ec-f9e067db0320) so the Monday /improve cron picks it up. Two concrete fixes proposed: (1) require per-article why/mapping that keys off the article's specific argument; (2) default the Related section to >=2 same-cohort cross-links (sibling articles in start_here_category) plus the Apr-15 anchor.

2. Single-cross-link failure mode. 7 of 23 (~30%) Commoncog entries had only 1 Related wikilink — all to the same Apr-15 anchor. Auto-fixed by adding 3 same-cohort cross-links each. Same root cause as #1: the template hard-coded one anchor link without including siblings. The Notion /improve task above addresses this.

3. Frontmatter discipline is excellent across the board. Every entry has date/type/source/author/tags. Newsletter-format entries have sponsored: false populated. members_only is consistently flagged on the Commoncog entries. Zero frontmatter remediation needed.

4. No copy-paste walls detected. Every summary is original prose. The Commoncog entries paraphrase rather than quote, and the deep-research briefs synthesize across multiple sources without lifting passages.

5. Conciseness is good. Largest non-research entry is [[synthesis-harness-thesis-dissent-2026-04-12|2026-04-12-harness-thesis-dissent]] at 86 lines. The three research briefs run 100-117 lines, justifiably dense for their scope. No bloat.

Process recommendations

  1. Apply the /improve task above on the next Monday cron — should land before any future cohort backfills.
  2. Hand-add 1-2 sibling cross-links to [[2026-04-19-commoncog-ultimate-guide-reading-book-a-week|commoncog-ultimate-guide-reading-book-a-week]] next pass (the lone B-grade entry).
  3. The deep-research-brief format is the current template-grade for the vault; consider pinning one as the canonical reference example in 02-sops/ for future briefs to imitate.

Review 3 — 2026-04-20

Reviewer: Ray (AI COO) Scope: 30 entries from 06-reference/ modified since 2026-04-13, excluding entries already scored in Review 2 and entries under transcripts/. Reviewed via /self-review --since 7d --limit 30 --fix. Max score: 13

Scored results

# File FM (2) Why (3) Map (3) Links (2) Bias (1) Walls (1) Concise (1) Total Grade
1 [[2026-04-20-practical-engineering-hidden-engineering-runways]] 2 3 3 2 1 1 1 13 A
2 [[2026-04-20-3blue1brown-volume-higher-dim-spheres-most-beautiful-formula]] 2 3 3 2 1 1 1 13 A
3 [[2026-04-20-indydevdan-pi-agent-teams-harness-engineering]] 2 3 3 2 1 1 1 13 A
4 [[2026-04-20-indydevdan-claude-code-2-0-agentic-coding]] 2 3 3 2 1 1 1 13 A
5 [[2026-04-20-indydevdan-top-5-agentic-bets-2026]] 2 3 3 2 1 1 1 13 A
6 [[2026-04-20-indydevdan-agent-experts-self-improving]] 2 3 3 2 1 1 1 13 A
7 [[2026-04-20-3blue1brown-exploration-epiphany-paul-dancstep]] 2 3 3 2 1 1 1 13 A
8 [[2026-04-20-3blue1brown-manim-demo-ben-sparks]] 2 3 3 2 1 1 1 13 A
9 [[2026-04-20-3blue1brown-grovers-algorithm-clarification]] 2 3 3 2 1 1 1 13 A
10 [[2026-04-20-indydevdan-agent-threads-boris-cherny]] 2 3 3 2 1 1 1 13 A
11 [[2026-04-20-indydevdan-one-agent-to-rule-them-all]] 2 3 3 2 1 1 1 13 A
12 [[2026-04-20-indydevdan-big-3-super-agent]] 2 3 3 2 1 1 1 13 A
13 [[2026-04-20-data-engineering-weekly-issue-266]] 2 3 3 2 1 1 1 13 A
14 [[2026-04-20-tim-ferriss-jamie-foxx-interview]] 2 3 3 2 1 1 1 13 A
15 [[2026-04-20-tim-ferriss-jordan-peterson-rules-psychedelics-bible]] 2 3 3 2 1 1 1 13 A
16 [[2026-04-20-tim-ferriss-personal-journaling-system]] 2 3 3 2 1 1 1 13 A
17 [[2026-04-19-tim-ferriss-eggs-without-peeling]] 2 3 3 2 1 1 1 13 A
18 [[2026-04-19-tim-ferriss-gabor-mate-trauma-addiction-ayahuasca]] 2 3 3 2 1 1 1 13 A
19 [[2026-04-19-tim-ferriss-gabor-mate-anger-rage]] 2 3 3 2 1 1 1 13 A
20 [[2026-04-19-hengsperger-reindustrialize-america]] 2 3 3 2 1 1 1 13 A
21 [[2026-04-19-tim-ferriss-huberman-foundations-physical-mental-performance]] 2 3 3 2 1 1 1 13 A
22 [[2026-04-19-tim-ferriss-jocko-willink-scariest-navy-seal]] 2 3 3 2 1 1 1 13 A
23 [[2026-04-19-tim-ferriss-healthy-breakfast-3-minutes]] 2 3 3 2 1 1 1 13 A
24 [[2026-04-19-tim-ferriss-brene-brown-save-your-marriage]] 2 3 3 2 1 1 1 13 A
25 [[2026-04-19-tim-ferriss-naval-ravikant-happiness-anxiety]] 2 3 3 2 1 1 1 13 A
26 [[2026-04-19-tim-ferriss-evening-routine]] 2 3 3 2 1 1 1 13 A
27 [[2026-04-19-tim-ferriss-how-to-remember-what-you-read]] 2 3 3 2 1 1 1 13 A
28 [[2026-04-19-tim-ferriss-how-to-use-writing-to-sharpen-thinking]] 2 3 3 2 1 1 1 13 A
29 [[2026-04-19-tim-ferriss-how-to-speed-read]] 2 3 3 2 1 1 1 13 A
30 [[2026-04-19-indydevdan-ditching-mcp-servers]] 2 3 3 2 1 1 1 13 A

No ⚠️ audit-failed annotations: zero overlap between the scored 30 and the 21 audit-failed files from the most recent audit (the audit-failed set is entirely from 2026-04-13 → 2026-04-18; the scored 30 are all 2026-04-19 → 2026-04-20).

Grade distribution

Grade Count %
A (12-13) 30 100%
B (10-11) 0 0%
C (8-9) 0 0%
D (<8) 0 0%

Summary statistics

Top entries (template-grade)

This entire batch is template-grade. The standout cluster: the 2026-04-20 IndyDevDan + 3Blue1Brown + Practical Engineering cohort (entries #1–12) all share the same structural pattern — explicit Why-in-vault with multi-paragraph rationale, 6-9 numbered Core argument items, 6-8 Mapping bullets each tied to a named RDCO surface, 5-7 Open follow-ups with time estimates, explicit Sponsorship section with bias-flagging discipline, and 5-7 wikilinks in Related. These are the new house style — propose pinning one (e.g. [[2026-04-20-indydevdan-agent-experts-self-improving]] or [[2026-04-20-practical-engineering-hidden-engineering-runways]]) as the canonical example in [[2026-04-19-newsletter-output-invariants|newsletter-output-invariants]].

The Tim Ferriss cluster (entries #14–28) is also exemplary across 15 entries despite high topical variance (kitchen hack, Navy SEAL, neuroscientist, vulnerability researcher, polymath performer). All hold the same shape; the synthesis-per-entry varies as the source warrants.

Worst entries (post-fix)

None. This is the first 30/30-perfect pass.

Double-signal entries (audit-failed AND self-review-flagged)

None. Zero overlap this cycle.

Cross-cycle audit remediation

Although the audit-failed cohort (21 files, 32 violations) does not overlap with the scored 30, this pass applied conservative mechanical fixes to that cohort per the --fix rules in SKILL.md:

I3 — missing newsletter_format / sponsored (2 fixes):

I12 — missing Curation section header on curation-format files (3 renames):

I11 — missing Sponsorship header when sponsored=true (4 additions/renames):

Left alone (per --fix conservative rules): I8 (Mapping content), I9 (Why-in-vault content), I10 (filename-sender mismatch — would break wikilinks). 11 of the 21 audit-failed files have only these non-mechanical violations and need manual remediation, not auto-fixes. List for the next manual pass:

Of these 15 still-failing files, 8 are I10 (filename mismatch) which is a Path-A choice not auto-fixable, and 7 require human-judgment Mapping/Why content (I8/I9) or format-classification review (I12 on hybrid/thought-leadership entries).

Systemic patterns

1. The hand-crafted assessment template has fully converged on the audit-invariant spec. Every one of the 30 scored entries follows the same six-section template (Why this is in the vault → Core argument → Sponsorship if applicable → Mapping against Ray Data Co → Open follow-ups → Related). The IndyDevDan/3B1B/Practical-Engineering 2026-04-20 cohort and the Tim Ferriss 2026-04-19 cohort are clearly the result of the same author/process — likely the /process-youtube skill output, possibly hand-curated by the founder. Whichever it is, the template discipline is now load-bearing and consistent enough to trust.

2. The audit-failure cohort lives in a different production lane. All 21 audit-failed files are from 2026-04-13 → 2026-04-18, predating the cohort the scored 30 came from. The audit-failed set is dominated by I10 (filename-sender mismatch, 6 occurrences) and I8 (missing Mapping section, 6 occurrences). These are characteristic of an earlier, less-disciplined /process-newsletter pass that did not consistently produce the Mapping section or align filename to canonical sender slug. The convergence in the 2026-04-19 / 2026-04-20 cohort suggests the underlying skill (or operator) corrected after the 2026-04-19 audit ran — possibly in response to the audit feedback itself. Worth confirming with the founder whether /process-newsletter was updated between Apr 16 and Apr 19.

3. Filename-sender mismatch (I10) is the largest remaining audit gap. 6 files have I10. Per the SKILL.md guidance ("renaming files breaks wikilinks; this is a Path-A choice, not auto-fixable"), this requires a deliberate one-time pass that renames the file, updates the audit log, and rewrites every wikilink that points to the old name. Worth doing in a single batch with a script rather than manually. Estimated: 1 hour for the 6 files in the audit log, plus any others discovered in vault.

4. The hybrid/thought-leadership format ambiguity around I12 is real. Several files (Stratechery memos, Every "You're the Manager Now") are flagged as newsletter_format: hybrid but their bodies look more like single-essay thought-leadership with no curated link list. The audit's I12 invariant ("curation-format files require Curation section") fires false-positive on these. Recommend either: (a) tightening the audit to skip I12 for non-curation formats, or (b) reclassifying these entries to thought-leadership instead of hybrid. The latter is closer to ground truth — these are not curation pieces.

5. Nothing concerning on bias / walls / conciseness across the scored 30. Every sponsored entry has an explicit ## Sponsorship section flagging the bias-to-watch. Zero copy-paste walls detected (every Core argument is paraphrased into the assessment voice). Most entries are 60-75 lines including frontmatter and Related — the longest (the 2026-04-20-tim-ferriss-jamie-foxx interview) runs 70 lines and the density is justified by the source's range.

Process recommendations

  1. Do the I10 batch-rename pass as a single dedicated cycle. 6 files in the current audit log + likely more in the broader vault. Build a small script that: (a) reads filename, (b) reads frontmatter source/author, (c) computes canonical slug, (d) renames file, (e) greps for wikilinks to old name and rewrites. Sub-1-hour total.
  2. Tighten the I12 audit invariant to skip newsletter_format: hybrid and newsletter_format: thought-leadership entries — they don't structurally need a Curation section. Alternatively, reclassify the 4-5 hybrid entries that are really single-essay pieces.
  3. Pin one of the 2026-04-20 IndyDevDan or Practical Engineering entries as the canonical reference example in [[2026-04-19-newsletter-output-invariants|newsletter-output-invariants]]. The template has converged; capture it as the explicit reference so future ingestion skills (or operators) inherit it by default.
  4. Investigate whether /process-newsletter or /process-youtube was updated between Apr 16 and Apr 19. The visible quality jump in the 2026-04-19 / 2026-04-20 cohort vs. the 2026-04-13 / 2026-04-14 cohort suggests an in-flight skill improvement. If so, capture what changed in the SKILL.md history so the lift is reproducible.
  5. No new /improve task this cycle — the cohort backfill task from Review 2 (Notion page id 347f7d49-36d1-81eb-b0ec-f9e067db0320) is still the active improvement target.

improve_processed: 2026-04-20

/improve autonomous run — 2026-04-20

Review 4 — 2026-04-23

Reviewer: Ray (AI COO) Scope: 30 most-recently-modified entries in 06-reference/ since 2026-04-16, transcripts excluded Args: --since 7d --limit 30 --fix

improve_processed: 2026-04-24

/improve autonomous run — 2026-04-24

Review 5 — 2026-04-24

Reviewer: Ray (AI COO) Scope: 30 entries from 06-reference/ modified in last 7d (window 2026-04-17 → 2026-04-24), --fix mode, audit-aware Max score: 13

Headline numbers

Grade distribution

Grade Count %
A (12-13) 28 93%
B (10-11) 1 3%
C (8-9) 1 3%
D (<8) 0 0%

Entries needing attention

  1. [[2026-04-21-tim-ferriss-90-days-black-belt]] — Score 8/13 (Grade C)

    • Missing: why-in-vault (0/3), cross-links (0/2; only one bare-path reference and a non-existent file pointer)
    • Mapping section explicitly self-flags as "weak" and notes the entry passed the Tier-1 promo-clip duration stopgap by 68 seconds (308s vs 240s threshold)
    • Manual review action: this is a genuine candidate for archival or merge into the parent Michelle Khare interview entry once that's filed. Don't spend cycles upgrading; either archive or delete.
    • Also surfaces a real skill-prompt drift note already captured in the entry: tighten the promo-clip duration threshold to 360s/480s OR accelerate the phrase-overlap dedup work tracked at Notion 348f7d49-36d1-81ac-b9d8-fad3ae60ce72.
  2. [[design-vocabulary-glossary]] — Score 11/13 (Grade B)

    • This is a glossary-shaped concept doc, not a newsletter ingestion. The Mapping/Why-in-vault criteria don't fit the format cleanly; the doc explains its own purpose in lines 14-17 (replace gestural design talk with category labels).
    • No fix applied. Counted as B because the criteria are designed for newsletter/article entries; flagging as a known mismatch rather than a real quality gap.

Top entries (template-grade)

  1. [[2026-04-23-unhobbling]] — 13/13. Cross-cluster anchor; dense cross-links; explicit RDCO implications priority list; clean why-this-note-exists framing.
  2. [[2026-04-23-moonshots-elon-cursor-bet-claude-kills-saas-openai-departures]] (post-fix) — 13/13. Strong RDCO mapping with 9 distinct angles, sponsor disclosure, and 14+ wikilinks.
  3. [[2026-04-23-generative-engine-optimization-geo]] — 13/13. Companion canonical-term doc; same-shape RDCO-implications structure as unhobbling.

Systemic patterns

Process / improve recommendations

Pre-existing structural items (carrying)

Author advisor note

This is the cleanest cohort to date. The vault's quality bar is now stable at A-grade with structural fixes confined to "did the why-in-vault section get written." No new drift detected. The two prior /improve tasks from Review 4 (X-article frontmatter contract, audit-log pruning) remain valid but should ship before another self-review cycle so we can measure their effect cleanly.

improve_processed: 2026-04-27

Review 6 — 2026-04-26

Reviewer: Ray (AI COO) Scope: 30 entries from 06-reference/ modified in last 7d (window 2026-04-19 → 2026-04-26), --fix mode, audit-aware Max score: 13

Headline numbers

Grade distribution (post-fix)

Grade Count %
A (12-13) 29 97%
B (10-11) 0 0%
C (8-9) 0 0%
D (<8) 1 3% (tim-ferriss-90-days-black-belt, archive candidate)

Entries needing attention

  1. [[2026-04-25-squarely-current-state-review]] — Pre-fix Score 4/12 (Grade D) ⚠️ audit-failed

    • Pre-fix issues: missing why-in-vault (0/3), missing mapping section (0/3), zero wikilinks (0/2)
    • Audit invariants failed: I3 (newsletter-format/sponsored fields — N/A for internal-review but audit doesn't know that), I8 (missing Mapping section), I9 (missing Why section), I10 (filename sender slug doesn't match internal-review source)
    • Fixed: added why-in-vault block, added Mapping section with 5 specific RDCO connections (design-system inheritance, MAC trademark-first sequencing, author-identity narrative, domain-mismatch as Teton-class drift, A+ Content patterns reuse), added 4 wikilinks to Squarely project hub + sister SC pieces.
    • Post-fix: 12/12 (A). The I3/I10 audit fails persist because the audit doesn't yet know type: reference + source: internal-review is a legitimate non-newsletter shape — flag for /improve to consider an audit-side carve-out for source: internal-review (similar to what's likely needed for source: synthesis).
  2. [[2026-04-24-wai-starship-never-returned-spacex-history]] — Pre-fix Score 10/13 (Grade B) ⚠️ audit-failed

    • Pre-fix issue: missing why-in-vault (0/3); had Episode summary but no explicit "Why this is in the vault" header
    • Audit invariants failed: I3 (newsletter-format missing — YouTube content; audit invariant should soften for content_type:podcast), I9 (no Why section)
    • Fixed: added why-in-vault section anchored to the iteration-cadence-as-moat candidate concept and the design-as-feedstock pattern, cross-linked to the Apr 21 Flight 12 sister episode.
    • Post-fix: 13/13 (A).
  3. [[2026-04-21-tim-ferriss-90-days-black-belt]] — Score 7/13 (Grade D) — NOT auto-fixed (archive candidate)

    • Same entry flagged in Review 5. Tier-1 promo clip (5 min) from a longer parent interview; self-flags as low-value in its own Mapping section. No audit failures (frontmatter is technically clean for a YouTube-clip entry).
    • Action: hold for archive when parent Michelle Khare episode is filed, OR resolve via the promo-clip-duration-threshold tightening Notion task 348f7d49-36d1-81ac-b9d8-fad3ae60ce72.

Top entries (template-grade)

  1. [[2026-04-21-practical-engineering-teton-dam-failure]] — 13/13. The template for failure-case-study entries: 4 dense Mapping sub-sections explicitly cross-linked to existing concept docs (binary-decision-around-continuous-probability, operational-definitions, layered-defense-architecture); sponsor block disclosed; 8 wikilinks; the pattern every other dam-failure or process-control entry should be measured against.
  2. [[2026-04-23-moonshots-elon-cursor-bet-claude-kills-saas-openai-departures]] — 13/13. Long-form podcast entry with 10 RDCO mapping bullets, 16 cross-links, sponsor block disclosed.
  3. [[2026-04-22-garry-tan-skillify-it-workflow]] — 13/13. X-article entry with full table comparison of RDCO state vs. Tan state across the 10-step skillify checklist; converts external content into a concrete RDCO upgrade roadmap.

Systemic patterns

Process / improve recommendations

Pre-existing structural items (carrying)

Author advisor note

Vault quality bar is holding at A-grade. The structural fixes are now consistent enough that the dominant noise-source is the audit log surfacing already-fixed structural drift (the audit-log-pruning /improve task closes this) and the audit not knowing about non-newsletter ingestion shapes (X articles, internal-review docs, synthesis docs). Both are knowable invariant-side fixes. The single new pattern worth the founder's attention is the internal-review/state-snapshot folder decision — pick a home before the next one lands and the mapping-section discipline either becomes mandatory by structure or unnecessary by category.

improve_processed: 2026-04-27

/improve autonomous run — 2026-04-27

Review 7 — 2026-05-03

Reviewer: Ray (AI COO) Scope: 30 entries created/modified 2026-04-26 → 2026-05-03 (excluding 06-reference/transcripts/) Window flag: --since 7d --limit 30 --fix Audit pre-failure set (in window, intersecting reviewed cohort): 16 files (per ~/.claude/state/newsletter-audit-log.md, 19 audit runs in window)

Top-line numbers (post-fix)

Files actually fixed this cycle

File Pre-fix Post-fix What changed
[[2026-04-29-blender-guru-donut-tutorial-part-1]] 8/13 C 13/13 A added author, added Why-in-vault, renamed Cross-references → Related with 3 new wikilinks
[[2026-04-30-backfill-discovery-practical-data-modeling]] 8/13 C 13/13 A added author, added Mapping section (4 ops-bullets)
[[2026-04-27-moonshots-sinclair-longevity-pill]] 9/13 C 12/13 A added Why-in-vault (3 specific load-bearing claims framed); copy-paste flag remains (long blockquote-style segments)
[[2026-04-27-indy-dev-dan-maximize-claude-code-subscription]] 9/13 C 12/13 A added Why-in-vault (Mac-mini OAuth-token operational stakes); copy-paste flag remains
[[2026-04-29-dwarkesh-reiner-pope-gpt5-claude-gemini-training]] 9/13 C 12/13 A added type: reference, added Why-in-vault; copy-paste flag remains (long blackboard segments)
[[2026-04-29-alphasignal-warp-open-source-zed-gemma]] 10/13 B 13/13 A added type: reference, added sponsored: false
[[2026-04-29-data-engineering-central-ai-changing-de-fast]] 11/13 B 13/13 A added newsletter_format, sponsored: true, sponsor_entity; renamed type: newsletter-announcement → reference per audit I4
[[2026-04-30-sanity-check-bet-architecture-audit]] 11/13 B 13/13 A added source + author fields

Double-signal entries (audit + self-review both flagged) this cycle

File Self-review issue Audit invariants Status
[[2026-04-29-blender-guru-donut-tutorial-part-1]] missing why + 0 wikilinks I3, I4, I9 FIXED
[[2026-04-30-backfill-discovery-practical-data-modeling]] missing mapping + missing author I3, I8, I9, I10 FIXED (I10 may persist — slug-vs-source mismatch is structural)
[[2026-04-27-moonshots-sinclair-longevity-pill]] missing why I3, I9 FIXED (I3 is podcast vs newsletter shape — known false-positive class)
[[2026-04-27-indy-dev-dan-maximize-claude-code-subscription]] missing why I3, I9 FIXED (same as above)
[[2026-04-29-dwarkesh-reiner-pope-gpt5-claude-gemini-training]] missing why + missing type I3, I4, I9 FIXED

Top entries (template-grade, 13/13)

  1. [[2026-04-30-mitohealth-founder-5-layer-agent-native-company-loop]] — concept doc; 13/13. Strong load-bearing-data + sharp founder mapping.
  2. [[2026-04-30-meta-ads-cli-agent-native-launch]] — 13/13. Tight, focused, good cross-link density.
  3. [[2026-05-02-khairallah-ai-automation-playbook]] — 13/13. Concise (72 lines), explicit RDCO mapping.
  4. [[2026-04-25-squarely-current-state-review]] — 13/13. The internal-review carve-out (shipped after Review 6) is now paying off — this entry passes cleanly.
  5. [[2026-04-30-stratechery-amazon-earnings-trainium-commodity]] — 13/13. Newsletter-format-perfect.

Systemic patterns

Process / improve recommendations

Author advisor note

Bar holds at A-grade (~83% A, 0% C in the content cohort). Two systemic items worth founder eye: (1) why-in-vault auto-enforcement at write-time would compress the manual fix queue substantially, and (2) the audit could grow a podcast/tutorial/interview carve-out with the same shape as the internal-review one shipped 2026-04-27. Neither is urgent — vault quality is high — but both are knowable, low-risk skill-side fixes.

Review 8 — 2026-05-04

Reviewer: Ray (AI COO) Scope: 14 entries newly created since Review 7 (2026-05-03 strategic-conversation outputs + 2026-05-04 DEW #268). Window 2026-04-27 → 2026-05-04, but de-duped against Review 7's already-fixed cohort. Window flag: --since 7d --limit 30 --fix Audit pre-failure set (in window, intersecting reviewed cohort): 6 files (per ~/.claude/state/newsletter-audit-log.md, run 2026-05-03T12:57:44 — 9 files audited, 6 failed)

Top-line numbers (post-fix)

Files actually fixed this cycle

File Pre-fix Post-fix What changed
[[2026-05-03-icbd-holdings-curative-ai-research]] 4/13 D 13/13 A added source + source_url + author frontmatter; added ## Why this is in the vault heading; added 4-bullet ## Mapping against Ray Data Co section; added ## Related with 4 wikilinks
[[2026-05-03-tampa-target-shortlist]] 7/13 D 12/13 A added ## Mapping against Ray Data Co section; added ## Related with 5 wikilinks. Conciseness fail (326 lines) accepted — same project-doc class as 04-finance/ carve-out
[[2026-05-03-app-store-submission-canonical-sources]] 10/13 B 13/13 A added source_url, author, newsletter_format to frontmatter; converted Related-section paths to wikilinks (3 added, including [[2026-04-25-squarely-current-state-review]])

Double-signal entries (audit + self-review both flagged) this cycle

File Self-review issue Audit invariants Status
[[2026-05-03-writewithai-claude-design-guidelines]] no Mapping (intentional skip-stub) + 0 wikilinks I8, I9, I11 NOT FIXED — see Systemic Pattern #1

Top entries (template-grade, 13/13)

  1. [[2026-05-03-shopify-eng-ucp-technical-architecture]] — clean technical-blog assessment with explicit Mapping subsections per RDCO surface; 7 wikilinks; concise (123 lines).
  2. [[2026-05-04-dataengineeringweekly-268-agents-replacing-search]] — newsletter-format-perfect with explicit Sponsorship section + curation breakdown; ties Doug-Turnbull retrieval finding to Tristan-Handy BI thesis as a generalization data point.
  3. [[2026-05-03-alphasignal-single-vs-multi-agent-systems]] — exemplar bias-flagging (separates Lambda paid placement from internal AlphaSignal workshop CTA).
  4. [[2026-05-03-sytaylor-ucp-merchant-owned-agentic-checkout]] — explicit ## ⚠️ Bias disclosure block calling out the Tempo-CEO forwarding channel; the founder explicitly flagged the bias and the doc records it.

Systemic patterns

  1. Skip-stub class needs rubric carve-out (NEW). [[2026-05-03-writewithai-claude-design-guidelines]] is a deliberate status: skipped stub — the founder filed the article-skip rationale itself rather than do a full assessment because the article is sales-funnel filler that adds nothing to RDCO design systems. The rubric currently scores it as C (no Mapping, no Links by design), and the audit flags it (I8/I9/I11). Both are false positives for skip-stub class. Recommend: add a status: skipped carve-out — when frontmatter has status: skipped, only score Why-in-vault (which becomes "Why this is skipped") and require nothing else. Mirrors the Review 6 internal-review and Review 7 podcast/tutorial carve-outs.
  2. Project-doc conciseness false positive (CARRY-OVER + EXTENDS). Review 7 flagged 04-finance/ for scope-exclusion. Tampa shortlist (01-projects/acquisitions/) and Physical-AI opportunity-map (01-projects/physical-ai-thesis/) hit the same class — multi-page project shortlists/opportunity-maps where >300 lines is the correct shape, not bloat. Recommend: extend the scope refinement to cover 01-projects/<dir>/<dated-doc>.md shortlists/opportunity-maps OR raise the conciseness threshold to 400 for type: opportunity-research and type: acquisition-research. Tampa post-fix passes the rubric in every other dimension; the conciseness fail is the only blocker to A-grade and it's structural to the doc's purpose.
  3. Why-in-vault enforcement (CARRY-OVER from Review 6 + 7). Did not regress this cycle — every new newsletter-shape entry has the heading. The Review 6 /improve task to enforce at write-time still hasn't shipped, but the manual discipline is holding for the small set of new entries. Continue tracking.
  4. Bias/sponsor flagging is improving (POSITIVE TREND). Three entries this cycle (heyrico, sytaylor, alphasignal-single-vs-multi-agent) have explicit ## ⚠️ Bias disclosure or ## ⚠️ Sponsorship headings beyond the frontmatter sponsored: true flag, with bias source named (Tempo-CEO forwarding channel; Lambda paid placement vs internal AlphaSignal workshop). This is exactly the discipline the rubric wants and is now becoming default behavior.
  5. Audit pre-failure set is shrinking on new content. 6 of 14 entries (43%) had any audit failure this cycle, down from 16 of 30 (53%) in Review 7 normalized to the 14-entry sample. I10 (filename-sender-mismatch) and I3 (missing newsletter_format on essay/X-article shape) remain the dominant patterns — both are structural issues with the audit's invariants, not the content. Same systemic recommendation as Review 7: extend audit carve-out to content_type: essay | company-research | reference-catalogue for the I3 newsletter-format check.

Process / improve recommendations

Author advisor note

Bar continues to hold at A-grade (93% A; only C-grade is a deliberate skip-stub). Trend line is +0.5 vs Review 7 — the manual discipline founder + Ray are running on new entries is sufficient even without the Review 6 /improve task shipping. The two NEW systemic items worth founder attention this cycle: (1) skip-stub rubric carve-out (cheap, ships in one /improve pass), and (2) extending the project-doc scope-exclusion to 01-projects/. Neither requires founder judgment beyond a yes/no — both are skill-side fixes Ray can ship if approved.

/improve autonomous run — 2026-05-04

Reviews processed: Review 7 (2026-05-03) + Review 8 (2026-05-04). Both marked improve_processed: 2026-05-04 via inline HTML comment under their headings.

Low-risk fixes applied (4):

  1. Audit script — podcast/tutorial/interview carve-out (I3, I4). Edited ~/.claude/scripts/audit-newsletter-outputs.py to add is_youtube_long_form flag (matches content_type: podcast|tutorial|interview) and skip both required-field check (I3) and type-equals-reference check (I4) for that class. Mirrors the 2026-04-27 internal-review carve-out pattern. (Review 7 systemic pattern.)
  2. Audit script — skip-stub bypass (I8, I9, I11). Same script: added is_skip_stub flag (matches status: skipped) and bypassed the Mapping-section, Why-in-vault, and Sponsorship-section checks for that class. (Review 8 systemic pattern #1.)
  3. Audit script — Changelog block. Documented all three carve-outs (2026-04-19, 2026-04-23, 2026-04-27, 2026-05-04) in the module docstring so future changes have provenance.
  4. Self-review SKILL.md — scope exclusions + skip-stub special pass class. Edited step 1 to exclude 04-finance/ (Review 7), exclude 01-projects/*/<dated>.md (Review 8), and treat status: skipped files as a special pass class (effective max 4/4: only Frontmatter + No copy-paste + Conciseness count). Added Changelog section documenting the change. (Review 7 + 8 systemic patterns combined.)

Files modified (3):

Structural changes queued (0 new): the why-in-vault auto-enforcement at /process-youtube write-time (Review 7 + 8 carry-over) is already on the Notion board as task 34ff7d4936d181cc891ff4323fdaff45 from the 2026-04-27 /improve cycle (https://www.notion.so/34ff7d4936d181cc891ff4323fdaff45 — Status: To Do, Priority: Medium, Owner: Both, Project: Ops). No re-queue needed.

No-ops: none. All four candidate low-risk fixes shipped. Audit script syntax-validated via py_compile.

Pre-flight audit signals:

Review 9 — 2026-05-08

Reviewer: Ray (AI COO) Scope: 30 most-recent entries in 06-reference/ mtime ≥ 2026-05-01 (full window 2026-05-01 → 2026-05-08, capped at --limit 30). Window flag: --since 7d --limit 30 --fix Audit pre-failure set (in 7d window): 47 distinct files across 12 audit runs since 2026-05-01T12:56:50; 22 of those 47 fall within the 30-file scored cohort.

Top-line numbers (post-fix)

Files actually fixed this cycle

File Pre-fix Post-fix What changed
[[2026-05-07-ship30for30-grow-linkedin-2026]] 8/13 C 13/13 A added date, type: reference, author to frontmatter; renamed format → newsletter_format; added ## Why this is in the vault body section. Audit invariants resolved: I3, I4, I9, I11 (sponsorship section already present in body).
[[2026-05-07-alphasignal-stanford-deep-learning-throttling-multiagent]] 7/13 D 13/13 A added source: AlphaSignal, author: Lior Sinclair (AlphaSignal), sponsored: false to frontmatter; renamed ## Why-in-vault → ## Why this is in the vault (canonical heading). Audit invariants resolved: I3, I9.
[[2026-05-07-writewithai-voc-landing-page-claude-code]] 2/13 D 13/13 A added author, sponsored: true, sponsor_entity to frontmatter; renamed format → newsletter_format; changed type: newsletter-reference → type: reference; added ## Why this is in the vault body section; renamed ## Tactical mapping to RDCO → ## Mapping against Ray Data Co; added ## Related body section with 5 wikilinks. Audit invariants resolved: I3, I4, I8, I9.
[[2026-05-05-wai-spacex-starship-flight-12-launch-date]] 9/13 C 12/13 A added ## Why this is in the vault body section threading the WAI cadence to feedback-loop-primacy thesis + Critical-Component discipline. Conciseness still passes (96 lines). The 1pt residual is the copy-paste-wall heuristic flagging the 200+-word Episode summary paragraph (false positive — original assessment, not pasted).

Auto-fix tally

Flagged for manual review (NOT auto-fixed)

Top entries (template-grade, 13/13)

  1. [[2026-05-06-osmani-cognitive-surrender]] — clean source assessment, full frontmatter, concrete RDCO mapping, multiple wikilinks.
  2. [[2026-05-06-alphasignal-anthropic-finance-gpt45-grok43]] — newsletter-format-perfect with bias flagging, sponsorship section, curation breakdown.
  3. [[2026-05-06-ship30for30-3-mistakes-marketers-ai]] — sibling of the audit-failed 3-outcomes-marketers-want-ai from same sender same week; this one passes everything cleanly. Useful template reference for the watch-mode subagent prompt.
  4. [[2026-05-05-innermost-loop-singularity-and-regulators]], [[2026-05-05-naval-find-simplest-thing]], [[2026-05-05-naval-good-products-hard-to-vary]], [[2026-05-05-naval-judgment-decisive-skill]], [[2026-05-05-naval-specific-knowledge]], [[2026-05-05-jorgenson-almanack-of-naval-ravikant]], [[2026-05-05-every-codex-native-apps]] — multiple 13/13 entries from the same Naval-corpus batch indicate the L5-thesis-validation work pipeline is producing template-quality output.

Systemic patterns

  1. Watch-mode subagent prompt drift on /process-newsletter (CONFIRMED CARRY-OVER, ESCALATING). The 2026-05-07 batch produced 4 of 7 files with multiple-invariant violations: writewithai-voc (I3+I4+I8+I9), ship30for30-grow-linkedin (I3+I4+I9+I11), ship30for30-3-outcomes (I8), alphasignal-stanford (I3+I9). The defect class is consistent: subagents are emitting format instead of newsletter_format, newsletter-reference / newsletter instead of type: reference, omitting author, omitting sponsored: true|false, and using non-canonical Why-section headings. This is the systemic pattern context-noted in this run's task input — confirmed across the full 7d window, not just one batch. Recommend an /improve pass to harden the watch-mode subagent prompt template: enforce canonical frontmatter field names + canonical body-section headings as a checklist the subagent must validate before writing. Today's pre-fix double-signal count was 4 (all 4 C/D entries were also audit-failed); we cleaned that to 0 with --fix, but the recurrence rate per batch is the underlying problem.
  2. Copy-paste-wall heuristic is over-detecting (NEW). My local scorer flagged 8 entries as copy-paste-wall based on a >80-word paragraph without inline structure. Sampling those (stratechery-joanna-stern-interview, naval-nothing-ever-happens-is-over, dec-ai-not-replacing-curious-developers, wai-spacex-starship-flight-12, tim-ferriss-naval-ravikant-1-2015, ship30for30-claude-code-marketing, practical-engineering-physics-behind-thumb-trick, innermost-loop-event-stream, stratechery-microsoft-apple-earnings) shows most are original prose summaries by the founder or by the /process-* skills, not literal copy-paste from source. The heuristic needs refinement: paragraphs starting with a section-header phrase ("Episode summary", "Source", "Core thesis") are common original-summary patterns and shouldn't trip the wall detector. Recommend skill-side: keep the heuristic as a flag-only signal, never auto-fix (already correct), but add a 1pt downgrade only when wall + sibling-article-paste-detection both fire. For Review 9 I treated copy-paste-wall flags as soft (entries with otherwise complete content + this single flag still rated A 12/13 with the 1pt deduction).
  3. Audit-failed cohort intersection rate is high but mostly resolvable on individual entries (POSITIVE). 22 of 30 entries (73%) had any audit failure since 2026-05-01, but the violations are concentrated in: I3 (newsletter_format / sponsored field) on Naval-corpus and book-class entries that aren't actually newsletters (they're book-derived snippets and tweetstorm assessments), and I10 (filename-sender-mismatch) on book-class entries (jorgenson-almanack-of-naval-ravikant source='book' / slug-head 'jorgenson'). The I3/I10 pattern on Naval-corpus is the same class as the 2026-04-27 internal-review carve-out and the 2026-05-04 podcast/tutorial/interview carve-out. Recommend extending the audit script's content-type carve-out to include content_type: book-excerpt | tweetstorm | aphorism-collection | corpus-index (or whatever the actual frontmatter type values are on these files) so the audit doesn't generate noise on a class that's structurally different from newsletter-shape content. This is a Review 7 + 8 carry-over pattern that's now hitting the Naval batch.
  4. Naval-corpus batch produced 7 of 22 A-grade entries (POSITIVE). Despite triggering audit I3 on every Naval entry (because the audit expects newsletter_format on type: reference), the self-review rubric scored them A because they have proper Mapping + Why-in-vault + Cross-links + concise. The audit is the false-positive signal here, not the self-review. This is what double-signal entries SHOULD look like — a single signal alone may indicate a noisy invariant, not a content problem. Reinforces Pattern #3.

Process / improve recommendations

Author advisor note

Avg dropped 0.7pts vs Review 8, but the cohort tripled in size and was less hand-curated — 30 entries spanning 7 days vs 14 in Review 8. The post-fix distribution (73% A, 27% B, 0 C/D) is still strong. The standout finding is confirmed: watch-mode subagent prompt drift is now a third-week-in-a-row pattern, and manual --fix is patching the same defect class repeatedly. This is the right shape for an /improve pass and qualifies as DECISION NEEDED for founder review.

improve_processed: 2026-05-08

/improve autonomous run — 2026-05-08

Review 10 — 2026-05-10

Reviewer: Ray (AI COO) Scope: 30 most-recently-modified entries in 06-reference/ (mtime > 2026-05-03) Window flag: --since 7d --limit 30 --fix

Top-line numbers (post-fix)

Files actually fixed this cycle

File Pre-fix Post-fix What changed
[[2026-05-10-harness-moat-two-layers-portability]] 4/13 D 13/13 A Added source, author, sponsored: false to frontmatter; added canonical ## Why this is in the vault and ## Mapping against Ray Data Co sections inside the existing concept doc. The ## Related section with valid wikilinks was already present.
[[2026-04-10-polydao-weather-markets-assessment]] 9/13 C 13/13 A Added sponsored: false to frontmatter; added ## Mapping against Ray Data Co section anchoring to PM1 baseline work, autoinv stack reuse, Sanity Check editorial pattern, and tracked-author signal.

Auto-fix tally

Flagged for manual review (NOT auto-fixed)

Top entries (template-grade, 13/13)

  1. [[2026-05-10-addy-osmani-agent-harness-engineering]] — direct-from-source assessment with full frontmatter, concrete RDCO mapping, 6 wikilinks. Anchor of this week's harness-engineering thesis cluster.
  2. [[2026-05-09-tobi-lutke-river-public-channel-agent]] — clean newsletter-shape note even though audit flagged I3 (newsletter_format) — semantic content is template-quality.
  3. [[2026-05-09-avedissian-loop-is-moat-robotics]] — same pattern, audit-failed but self-review A.
  4. [[2026-05-08-jaya-gupta-shape-as-moat]] — concept-adjacent reference with strong cross-cluster wikilinks.
  5. [[2026-05-08-stratechery-earning-spending-weekly]], [[2026-05-08-dan-farrelly-background-agents-orchestration]], [[2026-05-08-alphasignal-anthropic-claude-office-chrome-ai]] — newsletter batch from 2026-05-08 all at 13/13 on self-review despite I3 audit hits on a subset.

Systemic patterns

  1. I3 (newsletter_format / sponsored) on the David Perell + Tim Urban backfill batches (CARRY-OVER, MEDIUM). 16 of the 18 audit-failed cohort files are I3 hits on essay-class entries (David Perell's 8 essays + Tim Urban's 9 essays from 2026-05-08). These are essay-from-corpus-index files, not newsletter dispatches. The Review 9 /improve cycle added a content-type carve-out for book-excerpt | tweetstorm | aphorism-collection | corpus-index. Recommend extending the carve-out one more time to include essay or essay-collection for the Perell + Urban shape (essays harvested from a thinker's website rather than a newsletter delivery). This is the third backfill batch that's hit I3-noise without indicating a real semantic defect.
  2. I2 (YAML parse error) on 3 books-class entries (NEW, MEDIUM). [[2026-05-05-beck-tidy-first-2024|beck-tidy-first-2024]], [[2026-05-05-beck-tdd-by-example|beck-tdd-by-example]], [[2026-05-05-hughes-quickcheck-property-based-testing|hughes-quickcheck-property-based-testing]] all triggered I2 ("mapping values are not allowed here"). All three score 12/13 on self-review (pass everything except sponsor field for sponsored=false). Suggests the YAML in these files has a colon-in-string defect (likely a book subtitle with a colon or a quote glyph). Recommend a targeted manual sweep to close the YAML so they pass audit too. Low effort, high readability dividend.
  3. I11 (sponsorship-section-when-sponsored) on [[2026-05-09-data-engineering-central-cognitive-overload-ai-development|data-engineering-central-cognitive-overload-ai-development]] (ISOLATED, LOW). Single file flagged sponsored=true with no ## Sponsorship section in body. Self-review scored 12/13 (sponsor flag fail). One-off rather than systemic — the watch-mode subagent prompt hardening from the 2026-05-08 /improve run appears to have caught most of the class but missed this one entry from 2026-05-09. Worth a second prompt-tightening pass if the next batch shows the same shape.
  4. Concept-doc-class needs canonical heading discipline (NEW, LOW). The harness-moat concept doc was authored by me in this session and shipped without canonical ## Why this is in the vault + ## Mapping against Ray Data Co headings. I assumed the existing "## The question that prompted this" + "## Productizable read" body sections were sufficient — they're not, the rubric requires literal canonical headings. Recommend adding a concept-doc skill or skill-extension that templates the canonical sections at write-time.

Process / improve recommendations

Author advisor note

Avg up 0.7pts vs Review 9, post-fix distribution is 100% A. The cohort included 18 audit-failed files but every single one passed the self-review semantic rubric. This is the cleanest separation between "structural drift" (audit signal) and "semantic drift" (self-review signal) we've seen in 10 reviews — confirms the value of running both tools independently. The 2026-05-08 /improve pass on the watch-mode subagent prompt appears to have largely held, with the exception being on essay-class backfills that need their own audit carve-out.


/improve autonomous run — 2026-05-11

Reviews processed: Review 10 (2026-05-10). Marked improve_processed: 2026-05-11 above.

Low-risk fixes applied (1)

  1. Carve-out #1 (Review 10 systemic pattern #1) — audit-newsletter-outputs.py I3/I10 essay carve-out. Extended the existing is_source_corpus content-type bypass to include essay, essay-collection, and long-form-essay (the actual string written by /process-newsletter for the David Perell + Tim Urban backfill batches from 2026-05-08). This is the third carve-out in the same shape (2026-05-04 podcast/tutorial/interview; 2026-05-08 book-excerpt/tweetstorm/aphorism-collection/corpus-index; 2026-05-11 essay/essay-collection/long-form-essay). Comment + changelog entry added. Python AST parse verified clean.
    • File touched: ~/.claude/scripts/audit-newsletter-outputs.py
    • Verification: re-ran audit with --since 2026-05-04. All 16 Perell+Urban entries from 2026-05-08 now pass. Window-scoped pass-rate (since 2026-05-08): 36/40 = 90% (up from 22/40 = 55% before carve-out). Broader window (since 2026-05-04): 51/88 = 57.95% — remaining failures are non-essay-class structural issues outside this carve-out's scope.

Structural changes queued to Notion (2)

  1. Task: Manual sweep: 3 books-class vault entries hit audit I2 (YAML parse error) — Carve-out #2 from Review 10. Files: beck-tdd-by-example.md, beck-tidy-first-2024.md, hughes-quickcheck-property-based-testing.md. Likely colon-in-title YAML defect. Acceptance: 3 files pass audit OR audit I2 tolerance widened.

  2. Task: Extend Notion Research Backlog Source select to include deep-research-derivative (or formalize Notes-prefix pattern in /deep-research skill) — Notion schema gap surfaced by yesterday's /deep-research run. Two acceptable paths documented in the task notes.

Doctor / eval-mine signals (informational)

Meta-fix proposal (forwarded, not auto-applied)

Per the skill's "no one-off work" test: the C6 finding ("Referenced script missing or non-executable") fires repeatedly across the xcode-build-* family and remotion-to-hyperframes / spm-build-analysis. Same pattern observed in prior /improve cycles. Proposal: have /improve auto-flag a meta-fix when ≥3 skills reference the same missing script, OR widen rdco-doctor C6 to differentiate "script referenced but never authored" (likely vapor) from "script existed and got deleted" (regression). Not applied this cycle — flagging for a future /improve meta-loop pass on the doctor's heuristic.

Files touched

State


Review 11 — 2026-05-17

Reviewer: Ray (AI COO) Scope: 30 most-recently-modified entries in 06-reference/ (mtime > 2026-05-10) Window flag: --since 7d --limit 30 --fix

Top-line numbers (post-fix)

Files actually fixed this cycle

File Pre-fix Post-fix What changed
[[2026-05-16-moonshots-ep255-anthropic-spacex-leopold-singularity-economy]] 10/13 B 13/13 A Added canonical ## Why this is in the vault section drawn from existing Mapping content. Audit I9 cleared.
[[2026-05-10-tim-ferriss-cathy-lanier-rookie-cop-bricks]] 10/13 B 12/13 A Added canonical Why-heading explicitly framing this as a clip-companion to the parent interview note. Audit I9 cleared.
[[2026-05-14-tim-ferriss-tae-jin-park-prisoner-no-more]] 10/13 B 12/13 A Added canonical Why-heading flagging the file-and-forget intent. Audit I9 cleared.
[[2026-05-14-tim-ferriss-jerzy-gregorek-cerebral-palsy-coaching]] 10/13 B 12/13 A Added canonical Why-heading explicitly noting weak-mapping rationale. Audit I9 cleared.
[[2026-05-11-tim-ferriss-current-longevity-stack]] 10/13 B 13/13 A Added canonical Why-heading anchored to longevity-roster cross-check. Audit I9 cleared.
[[2026-05-09-tim-ferriss-most-ai-companies-wont-survive]] 10/13 B 13/13 A Added canonical Why-heading naming the four cluster connections. Not in pre-failure audit set (was already passing audit despite missing the heading — semantic-only fix).
[[2026-05-11-indy-dev-dan-delete-bash-tool-agentic-security]] 10/13 B 13/13 A Added canonical Why-heading. Audit I9 cleared.
[[2026-05-09-moonshots-ep254-google-record-quarter-white-house-gpt55]] 10/13 B 13/13 A Added canonical Why-heading. Semantic-only fix (not in audit-failure set).
[[2026-05-15-product-design-online-fusion-day-12-screwdriver]] 10/13 B 11/13 B Added canonical Why-heading (cleared audit I9). Frontmatter still YouTube-schema (no newsletter_format / sponsored) — manual carve-out work, NOT auto-fixable without a schema decision.
[[2026-05-15-dwarkesh-eric-jang-alphago-from-scratch]] 10/13 B 11/13 B Same — canonical Why-heading added (cleared audit I9). Frontmatter remains YouTube-schema.

Auto-fix tally

Flagged for manual review (NOT auto-fixed)

Top entries (template-grade, 13/13)

  1. [[2026-05-16-cfo-secrets-working-capital-warfare-iii]] — series-mode entry with explicit stacking on prior issues, sponsor block transparency, 5 named RDCO bet implications (Squarely / MAC / Sanity Check / RDCO holding-co / next-week's funding piece).
  2. [[2026-05-14-treybig-how-agents-use-systems-differently]] — investing-thesis-grade table mapping each named startup to the RDCO [[2026-04-24-targeting-system|targeting system]] pass/defer/yes filter, with explicit "where Treybig is light" honest-disclosure section.
  3. [[2026-05-14-mg-ecoatm-proposal-structural-patterns]] — 10-pattern structural absorption study with paired commercial+tech doc analysis, RDCO-application call-outs per pattern, and explicit "what RDCO should NOT borrow" section.
  4. [[2026-05-15-nateherk-3-ways-to-deploy-claude-agents]] — direct meta-architecture on RDCO's own deployment surface; surfaces a concrete experiment (two-loop /clear trick) queued for /improve.
  5. [[2026-05-15-every-team-agents-vs-personal-pets]] — clean customer-zero reading; 6 mapping points + same-day-cluster triangulation across [nateherk, agiledata, every-ai-work] sibling notes.

Systemic patterns

  1. I9 (missing canonical ## Why this is in the vault heading) on long-form podcast/interview entries (NEW, MEDIUM). 9 of 30 entries (30%) shipped without the canonical Why-heading despite having substantive why-content elsewhere in the body. All 5 Tim Ferriss YouTube entries, 1 IndyDevDan tutorial, 2 Moonshots episodes, and the (audit-passing) Elad Gil interview all jumped from # Title straight to ## Episode summary instead of ## Why this is in the vault → ## Episode summary. This isn't a content problem — every fixed entry had clear, non-generic why-content I could surface from the existing Mapping section. It's a template-discipline problem in the /process-youtube skill (and possibly the manual newsletter intake when source is a long-form YouTube interview). Recommend /improve extend the /process-youtube skill (and any related podcast/interview ingestion shape) to emit ## Why this is in the vault between # Title and ## Episode summary at write time. Concrete and tractable.

  2. YouTube non-standard frontmatter schema mismatch (CARRY-OVER FROM PRIOR REVIEWS, MEDIUM). Two files in window (product-design-online-fusion-day-12-screwdriver, dwarkesh-eric-jang-alphago-from-scratch) use /process-youtube watch Mode 4 default schema — frontmatter starts with source: youtube and lacks newsletter_format / sponsored / type: reference. Audit fires I3/I4/I9/I10 on all of them. Five Tim Ferriss YouTube files in the same cohort use a slightly different shape (source: Tim Ferriss (YouTube) + type: reference + content_type: interview) that mostly passes audit. This is the 4th distinct content-type-carve-out shape to surface across reviews — the prior three were 2026-05-04 podcast/tutorial/interview, 2026-05-08 book-excerpt/tweetstorm/aphorism-collection/corpus-index, and 2026-05-11 essay/essay-collection/long-form-essay. Recommend /improve either (a) add a 4th carve-out for source: youtube shape in the audit script, OR (b) tighten /process-youtube skill to always emit the audit-compatible shape (canonical date/type: reference/author/sponsored: false/newsletter_format: thought-leadership-or-equivalent at write time). (b) is the cleaner long-run fix because it converges the shapes; (a) is the immediate audit-noise relief.

  3. cfosecrets newsletter_format: mailbag not in audit enum (NEW, LOW). [[2026-05-12-cfosecrets-working-capital-design-growth]] (one entry, just inside the window-edge of the audit cohort but JUST outside the 30-file self-review cap) hit I5 with newsletter_format='mailbag'. The CFO Secrets shape sometimes runs Q&A-from-readers issues; mailbag is a legitimate newsletter-format category that the audit enum doesn't recognize. Either expand the audit enum to include mailbag OR rename to thought-leadership (it's borderline — Q&A is closer to thought-leadership than curation). Logging for /improve.

  4. The I8 / I12 / I10 hits on backfill-discovery + founder-synthesis + X-long-form files (LARGELY OUT-OF-COHORT, INFORMATIONAL). The 2026-05-17 audit run flagged 22 files; the self-review cohort captured 30 most-recent files. Of the 22 audit-failed files in the audit's broader 7d window, ~12 are inside our 30-file cohort and the rest are out-of-window (e.g. backfill-discovery-mostlymetrics, jaynitx, stripe-atlas-perks, zach-lloyd-warp, fde-wave-convergence, zack-igclaims, zoharatkins, the various cfosecrets-greatest-hits / dataengineeringcentral curation-format-missing-curation-section files). Pattern shape is the same as Review 10's call-outs — non-newsletter source shapes (X long-form, founder synthesis, discovery scans) tripping I3/I8/I10. The Review-10 essay carve-out helped; the 2026-05-17 set suggests one more carve-out shape pass on X-long-form + founder-synthesis would close most of the remaining audit-noise. Not urgent because the semantic content is template-quality on every one of those out-of-cohort files I sampled.

Process / improve recommendations

Author advisor note

Average score essentially flat from Review 10 (12.73 vs 12.63), but the shape of the cohort changed meaningfully. Prior weeks had no canonical-heading misses on long-form podcast/interview content; this week 9 of 30 entries (30%) shipped without canonical Why-heading despite having substantive content. The pre-fix B count was the highest in 4 review cycles, but the post-fix A count is 93%. The convergence holds: the audit (deterministic structural check) and self-review (semantic check) continue to flag different things, and the value of running both is unchanged.

Two systemic patterns this week are both about template discipline at write-time, not about content quality drift: (a) canonical Why-heading on long-form YouTube/podcast entries, (b) YouTube schema mismatch with audit invariants. Both are eminently fixable in the /process-youtube skill itself rather than via post-hoc audit carve-outs. Recommend /improve prioritize the skill-side fix (#1 above) over the audit-side carve-out (#2) because it converges shape rather than expanding tolerance.

The double-signal count dropped from 0 in Review 10 to 2 in Review 11 — both are the YouTube-schema mismatch files. Without the YouTube-skill fix, this count will likely grow as YouTube ingestion volume increases.

improve_processed: 2026-05-18

/improve autonomous run — 2026-05-18

Also applied (out-of-band, from rdco-doctor C6 finding same run):

Notes:


Review 12 — 2026-05-21 (on-demand kick-off; weekly cron scheduled Sundays 7:43am ET)

Inputs

Scoring distribution (post-fix)

Trend vs prior reviews

Fixes applied (--fix mode, low-risk template/schema corrections)

  1. [[2026-05-20-every-google-io-agents-anthropic-acquires-figma-vibe-check]] — frontmatter field-name drift. Pre-fix used non-canonical format: (vs newsletter_format:), sponsor: (vs sponsored: + sponsor_entity:), url: (vs source_url:), and was missing type: reference. Rewrote frontmatter to canonical schema. Content body (Why, Sponsorship, Mapping, Cross-refs) was already in shape — only the YAML field names were drifted.

  2. [[2026-05-20-alphasignal-gemini-omni-flash-antigravity-spec-kit]] — two issues fixed: (a) format: curation → newsletter_format: curation, (b) missing ## Why this is in the vault canonical heading despite having why-content in the body. Hoisted the existing intro paragraph under the canonical heading and added one sentence of RDCO-specific framing.

Audit ∩ self-review (double-signal entries)

Pre-fix: 2 (the same two files above) Post-fix: 0

Top files (template reference)

  1. [[2026-05-20-spacex-s1-ipo-filing-with-xai-consolidated]] — 13/13. SEC EDGAR primary-source + 12 wikilinks + multi-thesis mapping (xAI/Sanity Check angle, IPO-as-Anthropic-fundraise-evidence, dual-class structure).
  2. [[2026-05-18-acquired-vanguard-bogle-communist-capitalist]] — 13/13. Cleanest counterpositioning case-study filed in window; Sanity Check tie + RDCO-holding-co tie + 5 wikilinks.
  3. [[2026-05-20-dataengineeringcentral-ben-rogojan-left-facebook-podcast]] — 13/13. Triangulation pair (Jeff exit-debrief + Ben Rogojan + Alex Vacca services-as-software) is exactly the multi-source pattern reviews 9-11 surfaced as A-grade.

Systemic patterns

  1. format: vs newsletter_format: field-name drift (NEW, MEDIUM). 2 of 30 files (7%) used the bare format: key. Both were Every (jack-cheng article) and AlphaSignal (Lior daily roundup). Both were processed via /process-newsletter watch mode in the last 24h. The two source shapes converge on the same wrong field name, which suggests either (a) the watch-mode sub-agent prompt has the wrong field-name in its template, OR (b) two ingest paths quietly drifted simultaneously. Recommend /improve audit the /process-newsletter watch-mode sub-agent dispatch template and confirm it emits newsletter_format: not format:. Tractable, single-file fix.

  2. /process-youtube skill fix from Review 11 is HOLDING (CONFIRMATION, GOOD-NEWS). 5 YouTube files this cohort (Mosul Dam, Tim Ferriss Cathy Lanier x2, IndyDevDan, Acquired Vanguard) all emit the canonical ## Why this is in the vault heading between # Title and ## Episode summary. Review 11's systemic pattern #1 (30% miss rate on canonical Why heading for YouTube/podcast) is fixed at the skill level. No regressions in this cohort.

  3. YouTube schema audit-noise CONTINUES (CARRY-OVER from Review 11 pattern #2, LOW). The 5 YouTube files still don't have newsletter_format: in frontmatter — they use content_type: tutorial/podcast/interview. The /improve cycle on 2026-05-18 added a content_type: talk carve-out to the audit script but the broader carve-out (skip newsletter_format enforcement when source is YouTube) is still partial. Self-review treats these as A (the content is template-grade); the audit-script trip is noise. No further action requested — the fix path is correct, just incomplete.

Surfaced for /improve next cycle

Author advisor note

Clean week. The two B-grade files were both fixed in this run via deterministic field-name corrections — no semantic content rewrites needed, the Why/Mapping/Cross-refs were already in place. Mean score back to 13.0 after the 12.73 dip in Review 11. The Review 11 /process-youtube skill fix is fully holding — that's the load-bearing observation from this cycle. Single new systemic pattern (#1, the format: vs newsletter_format: drift on two /process-newsletter watch outputs) is the only thing worth /improve attention next cycle. Decision flag: none — clean week.

improve_processed: 2026-06-01


Review 13 — 2026-05-24 (weekly cron, Sunday 07:43 ET)

Inputs

Scoring distribution (post-fix)

Trend vs prior reviews

Fixes applied (--fix mode, low-risk schema/disclosure corrections)

File Fix Audit invariant cleared
[[2026-05-22-aparente-gist-tufte-viz-skill]] Added newsletter_format: thought-leadership to frontmatter (gist-class non-newsletter source needed the field for audit) I3
[[2026-05-21-mostlymetrics-spacex-ipo-s1-breakdown]] Changed newsletter_format: single-thread deep dive → thought-leadership (free-text value not in audit enum) I5
[[2026-05-21-innermost-loop-may-21-openai-erdos-disproved]] Changed newsletter_format: essay → thought-leadership (Wissner-Gross essay-shape; closest match in enum) I5
[[2026-05-22-tim-ferriss-sami-inkinen-virta-t2d-rowing]] Added sponsor_entity field + new ## Sponsorship section to body (sponsored=true had no disclosure block) I11
[[2026-05-23-moonshots-ep-257-spacex-ipo-gpt55-erdos]] Renamed body section ## Sponsors → ## Sponsorship, added sponsor_entity field + investor-positioning disclosure (Diamandis xAI investor, Wissner-Gross security plug) I11
[[2026-05-22-benn-stancil-wac-wins-above-claude]] Changed source: The Bridge → source: Benn Stancil (The Bridge) so filename slug-head benn-stancil appears in source string I10

Total: 6 files. All 6 audit-failed double-signal files cleared by single-file edits. No mapping rewrites, no copy-paste fixes (none found), no thin-content archiving.

Audit ∩ self-review (double-signal entries)

Top files (template reference, 13/13)

  1. [[2026-05-23-nick-prince-spacex-ipo-agent-ic-memo-x402]] — third agent-architecture piece in 24h (after Tony Dang Infisical, Harrison Chase LangSmith); sponsor disclosure with explicit dual-framing (real-analysis vs demo-for-x402); 6 wikilinks; direct paper-trade-feed mapping for elon-verse-v1.
  2. [[2026-05-21-mostlymetrics-spacex-ipo-s1-breakdown]] — CJ Gustafson SaaS-CFO lens enumerating 10 new disclosures yesterday's EDGAR-direct vault note missed (Mars-colony comp gates, Valor $20.2B guarantee, Cursor $10B termination, 30% retail directed-share dynamics). Direct thesis-update sweep trigger. 4 wikilinks. 168 lines, dense.
  3. [[2026-05-22-austin-vernon-american-manufacturing-essay]] — Missing-Middle thesis; founder reshoring-interest anchor. 4-shape decision frame (investment / employer-pivot / advisory / Sanity Check), explicit Hengsperger-vs-Vernon contrast. Live cross-link to SendCutSend $110M validation in the same week's Not Boring issue.
  4. [[2026-05-23-cfosecrets-working-capital-warfare-iv-funding-the-cycle]] — series mode entry (part 4 of 5); NEW Stuut sponsor cataloged; 8-funding-flavor taxonomy preserved; series-craft note for Sanity Check borrow. 5 wikilinks across series + adjacent.
  5. [[2026-05-20-shannholmberg-hermes-agent-control-room-four-levels]] — clean port-delta analysis (what RDCO has / what's missing / 4 tactics to port). Names control-plane-vs-runtime split RDCO operates on implicitly. 5 wikilinks.

Systemic patterns

  1. newsletter_format enum drift on essay-shape and deep-dive shapes (NEW, LOW-MEDIUM). 2 of 30 files (7%) used free-text values (essay, single-thread deep dive) that aren't in the audit's allowed enum. Both are intelligent author-choices (Wissner-Gross essays ARE essays; CJ's SpaceX S-1 breakdown IS a single-thread deep dive) but the audit enum is fixed at 7 values (business-history, curation, founder-interview, guest-post, hybrid, mailbag, thought-leadership). The skill is forced to compress meaningfully-distinct formats into thought-leadership as the catch-all. Two options for /improve:

    • (a) Expand the enum to include essay, single-thread-deep-dive (separate from thought-leadership which covers shorter takes)
    • (b) Hold the enum fixed and update /process-newsletter skill to map essay→thought-leadership at write-time Recommend (a) because the distinction is informationally useful (single-thread-deep-dive is a meaningfully different content shape than thought-leadership) and the audit-script enum is cheap to extend. This is the THIRD newsletter_format enum-drift surfaced across reviews (Review 11 mailbag, Review 12 format:-field-name-drift, Review 13 essay/single-thread). The pattern points at the enum being under-specified, not at the skill writing wrong values.
  2. I11 (sponsored=true with no ## Sponsorship body section) on long-form podcast/interview YouTube entries (NEW, MEDIUM). 2 of 30 files (7%) had sponsored: true in frontmatter but no canonical ## Sponsorship body block. Both were YouTube-source long-form interviews/podcasts (Tim Ferriss Sami Inkinen + Moonshots ep 257). The Moonshots file had a ## Sponsors (plural) section with the right content — just wrong heading. The Tim Ferriss file had no sponsorship body coverage at all despite frontmatter sponsored: true. Both processed via /process-youtube (Mode 4 watch fan-out). Recommend /improve extend /process-youtube skill to emit canonical ## Sponsorship (singular, exact wording) heading when sponsored: true, mirroring the same canonical-heading discipline Review 11 fixed for ## Why this is in the vault. Adjacent fix path: also enumerate at least the standard sponsor categories in the body (host-own-clinic conflicts, recurring-rotation-sponsors, investor-positioning) so the disclosure is non-empty.

  3. /process-newsletter field-name drift from Review 12 NOT recurring this cohort (CONFIRMATION). Review 12 flagged format: vs newsletter_format: on 2 files. Zero format: field-name drift in Review 13 cohort. Either /improve fixed the dispatch template, or the failure mode was transient. Worth one more review to confirm hold.

  4. I10 (filename-source-mismatch) on author-named source slugs (NEW, LOW). 1 file (benn-stancil-wac) had filename starting with author-name (benn-stancil-wac-...) but source field = publication name only (The Bridge). Author was in author: field. Fix landed by changing source to parenthetical Benn Stancil (The Bridge). Pattern to watch: when filename slug-head is the author-name (not the publication-name), source field should either lead with author-name or include it parenthetically. Consider extending audit-script tolerance to also check author: field for slug-head match — would avoid forcing the source field to absorb the disambiguation. Low priority because pattern is rare.

Surfaced for /improve next cycle

Author advisor note

Two consecutive 100%-A weeks (Reviews 12 + 13). The semantic-quality floor is now stable — the variance has compressed entirely into schema/disclosure-tier issues that the audit catches and --fix can resolve in single-file edits without semantic rewrites. This is the convergence pattern Reviews 7-11 were grinding toward.

Two of the three new systemic patterns this week (enum expansion, canonical-Sponsorship-heading) are skill-side fixes that would prevent the failure from recurring in future cohorts. Pattern #1 is the most leverage — the newsletter_format enum has now been surfaced for expansion three reviews in a row (mailbag, format-field-name, essay/single-thread), and the enum-vs-skill-template tradeoff has tilted toward expand-enum because the alternative (force every author-chosen format string into 7 buckets) loses information without buying anything.

Decision flag: none. Clean week. The Tim Ferriss YouTube/podcast I11 hit is the only entry where the founder might want to spot-check that the canonical sponsorship section I added actually reads as he'd expect (it's a generic recurring-sponsor disclosure since the Tim Ferriss show's standard sponsor rotation wasn't enumerated in the underlying transcript). If he wants tighter per-episode sponsor enumeration, that's a /process-youtube skill change, not a self-review change.

improve_processed: 2026-05-25

/improve autonomous run — 2026-05-25

Two input sources this run: (A) ~/.claude/state/improve-queue.md (6 findings from 2026-05-24 cron runs) and (B) review-log.md Review 13 (2026-05-24, 2 systemic patterns). De-duped: queue items 2+3 overlap exactly with Review 13 patterns #1+#2 — applied once, credited both sources.

Step 0 audits (folded in, no new actionable findings):

Reviews processed: Review 13 (2026-05-24). Marked improve_processed: 2026-05-25 above. Plus all 6 improve-queue.md items (queue rewritten: items 1–5 removed as applied, item 6 left with QUEUED pointer to Notion).

Low-risk fixes applied: 6 edits across 4 files (5 of 6 queue items)

Structural changes queued: 1 (queue #6 — the judgment call)

No-ops (unclear or already-fixed patterns): 3

Silent run on the low-risk side (changelog + this report are the audit trail). One structural change queued, so a one-line note posts to Discord #ops per the SKILL.md Step 6 protocol.

2026-05-31 (Review — autonomous, --since 7d --limit 30 --fix)

improve_processed: 2026-06-01

/improve autonomous run — 2026-06-01

2026-06-07 (Review — direct /self-review --since 7d --limit 30 --fix)

/improve autonomous run — 2026-06-08

Review 15 — 2026-06-14

Reviewer: Ray (AI COO) Scope: entries modified 2026-06-07 → 2026-06-14; 7 from 06-reference/, 2 from 08-tooling/; 04-finance/ and 06-reference/transcripts/ excluded per scope rules Args: --since 7d --limit 30 --fix Max score: 13

Headline numbers

Scored results

# File FM (2) Why (3) Map (3) Links (2) Bias (1) Walls (1) Concise (1) Pre-fix Post-fix Grade
1 [[2026-06-13-innermost-loop-export-control-singularity-curation]] ⚠️ audit-failed 2 3 3 2 1 1 1 13 13 A
2 [[2026-06-12-stratechery-twis-hey-siri-fable]] ⚠️ audit-failed 2 3 3 2 1 1 1 13 13 A
3 [[2026-06-10-every-how-to-get-the-most-out-of-fable-5]] 2 3 3 2 1 1 1 13 13 A
4 [[2026-06-10-shopify-engineering-quick-internal-hosting]] 2 3 3 2 1 1 1 13 13 A
5 [[2026-06-09-claude-fable-5-mythos-5-release]] 2 3 3 2 1 1 1 13 13 A
6 [[2026-06-09-mostly-metrics-will-fpa-eat-ir]] 2 3 3 2 1 1 1 13 13 A
7 [[2026-06-07-mostly-metrics-google-80b-equity-raise]] 2 3 3 2 1 1 1 13 13 A
8 [[2026-06-13-open-knowledge-format-okf-assessment]] 2 3 3 0→2 1 1 1 11 B 13 A (post-fix)
9 [[2026-06-09-fable5-workflow-optimization-memo]] 2 0→3 3 0→2 1 1 1 8 C 13 A (post-fix)

Files fixed this cycle

File Pre-fix Post-fix What changed
[[2026-06-13-innermost-loop-export-control-singularity-curation]] ⚠️ 13/13 A 13/13 A Added ## Curation section header before curated-items bullet list (clears audit I12)
[[2026-06-12-stratechery-twis-hey-siri-fable]] ⚠️ 13/13 A 13/13 A Renamed ## Digest highlights → ## Curation section (clears audit I12)
[[2026-06-13-open-knowledge-format-okf-assessment]] 11/13 B 13/13 A Added ## Related section with 3 wikilinks (2pt cross-links fix)
[[2026-06-09-fable5-workflow-optimization-memo]] 8/13 C 13/13 A Added ## Why this is in the vault (3pt) + ## Related section with 3 wikilinks (2pt)

Audit-failed entries (structural only, no semantic deficit)

  1. [[2026-06-13-innermost-loop-export-control-singularity-curation|innermost-loop-export-control-singularity-curation]] — I12: newsletter_format: hybrid lacked ## Curation section. Hybrid newsletter where curated items appeared under ## Summary without a separate Curation header. Fixed: added canonical header.
  2. [[2026-06-12-stratechery-twis-hey-siri-fable|stratechery-twis-hey-siri-fable]] — I12: newsletter_format: curation lacked ## Curation section. Weekly digest used ## Digest highlights instead. Fixed: renamed.

Both score 13/13 semantically. Zero double-signal entries.

Systemic patterns

  1. I12 header drift persists (4th cycle, same fix each time). ## Digest highlights / ## Summary / ## Issue contents variants keep appearing instead of the canonical ## Curation section. Every occurrence is a mechanical rename — no semantic failure underneath. Root cause is authoring habit on the non-watch path (or SKILL.md not prescribing the exact canonical header string). The /improve Notion task 379f7d4936d181589999ea86a264655d (canonical 06-reference filing contract for the non-watch path) covers this. Will recur until that task ships a write-time enforcement rule.

  2. Tooling memos lack structural scaffolding (Why-in-vault, Related). Both 08-tooling entries reviewed this week were missing these sections — not because the content was thin, but because self-generated memos don't go through the same write-time checklist as newsletter notes. The compile-vault SKILL.md fix from the 2026-06-08 /improve run added the checklist for concept articles; tooling memos are not yet covered.

  3. Newsletter + content ingestion pipeline is clean. All 7 06-reference/ notes are perfect-score. The watch-path enforcement is working; the non-watch path continues to be the quality gap.

Top entries (template-grade)

  1. [[2026-06-09-mostly-metrics-will-fpa-eat-ir]] — 13/13. Exemplary dual-sponsor disclosure (Brex as third-party + Mostly Talent as self-consulting arm). Why-in-vault is specific to the role-convergence thesis analog; Mapping identifies the meta-level insight (title holds, content merges) as external evidence for RDCO's own targeting-systems thesis.
  2. [[2026-06-10-every-how-to-get-the-most-out-of-fable-5]] — 13/13. Operational doctrine written by the model that upgraded the same day. Four worked examples mapped to specific RDCO harness primitives; pricing cliff corroborated. The dispatch-gate reframing of the four-quality task filter is the highest-value mapping bullet.

Process / improve recommendations

Pre-existing structural items (carrying)

improve_processed: 2026-06-15

/improve autonomous run — 2026-06-15

/improve autonomous run — 2026-06-22

Review 16 — 2026-06-28

/improve autonomous run — 2026-06-29

Review 17 — 2026-07-05

/improve autonomous run — 2026-07-06

Review 18 — 2026-07-12

/improve autonomous run — 2026-07-13

Review 19 — 2026-07-19

improve_processed: 2026-07-20

/improve autonomous run — 2026-07-20

Review 20 — 2026-07-26

improve_processed: 2026-07-27

/improve autonomous run — 2026-07-27

Review 21 — 2026-08-02

Scope: 30 entries across 06-reference/ (incl. concepts/ and research/), 02-sops/, 03-contacts/, 08-tooling/ — the 30 newest by mtime in [2026-07-26, 2026-08-02], 01-projects/* dated shortlists and 04-finance//06-reference/transcripts/ excluded per scope. --since 7d --limit 30 --fix. Scoring delegated to 5 parallel fresh-eyes subagents (6 files each, no shared context), each independently verifying every wikilink target against the filesystem via find.

improve_processed: 2026-08-03

/improve autonomous run — 2026-08-03

2026-08-09 (Review 22)

Scope: 30 entries across 06-reference/ (incl. 06-reference/research/) and 08-tooling/, dated 2026-08-01 → 2026-08-08 (--since 7d --limit 30 --fix). Reviewed as 3 parallel batches of 10; this entry consolidates all three.

improve_processed: 2026-08-10

/improve autonomous run — 2026-08-10

Low-risk fixes applied: 4 (2 skills, 2 scripts)

  1. self-review/SKILL.md — step 1.5 supersession rule. Review 22 reported [[2026-08-07-not-boring-wdoo-205]] as I11 audit-failed, observed the file carries a valid ## ⚠️ Sponsorship section, and concluded the audit script's header match was too strict — recommending the I11 regex be widened. Both halves are wrong. H_SPONSOR is ^##\s+(?:[^\w\s]+\s*)?Sponsorship\b, which already accepts the emoji-prefixed form (verified by direct match test against the actual file line, repr confirmed). And the audit log shows the file failing at 2026-08-07T12:04:30 and passing at 12:04:37 — a normal write → audit → fix → re-audit cycle. The invariant worked exactly as designed. The real defect is upstream: /self-review step 1.5 builds its audit pre-failure set from any logged violation without checking for a later clean run. Censused: 156 of 1015 logged failures (15.4%) are superseded — a material stale-signal rate feeding the double-signal report every cycle. Step 1.5 now requires a supersession check, and notes that clean runs report All audited files pass all invariants. without naming files, so the match must be on the run window, not a pass list.
  2. improve/SKILL.md — Recurrence guardrail check (7), now seven checks. A logged failure is pre-fix state, not live status. Append-only logs record run-time truth; audit→fix→re-audit means a cited violation is often already repaired. Check for a later covering run before acting — and when a review's story requires a tool to be broken, test the tool directly before believing it. Mirror of check 4 (that one: a repaired artifact doesn't falsify a finding; this one: a logged failure doesn't establish one).
  3. rdco-doctor.py — C6 false positives eliminated (2 → 0), after 3 cycles of being recorded as a stable no-op. Neither finding was real. (a) deep-research: build-research-digest.py — the skill runs cd ~/rdco-hq && python3 scripts/build-research-digest.py; the script exists at ~/rdco-hq/scripts/, but script_exists() only ever looked in ~/.claude/scripts/. Bare scripts/ refs now resolve against known repo script dirs. (b) investing-edgar-watch: edgar-fetch.py — the SKILL.md line literally reads "Planned location: … (not yet implemented — design stub only)"; refs on lines carrying a stub marker are now skipped. A residual third false positive surfaced once (a) landed — the digest script isn't chmod +x, but it is invoked through python3, so the bit is irrelevant; the exec check now applies only to directly-invoked refs. Deliberately did not chmod inside ~/rdco-hq, which would dirty the founder's working tree. Regression-tested: a genuinely missing script is still detected; skills/dark/overlap counts unchanged (57/40/1).
  4. eval-mine.py — the 4-cycle phantom signal is gone. Prior runs recorded "n=1 same false positive" three cycles running and declined to act. Root-caused this run: the 16 "still wrong" hits were one sentence of vault-health SKILL.md prose, injected as a user-role turn and re-counted once per session that loaded it — not 16 complaints. The existing guard's intent was right ("skip skill-body content embedded as user-role messages") but its heuristic was a 5000-char length threshold, and the injected body is 3421 chars. Replaced with the deterministic literal marker every injected skill turn carries (Base directory for this skill:). Miner now reports 0 hits / clean across 456 sessions — the truthful answer instead of four cycles of noise.

Structural changes queued: 0

No-ops

Process notes

2026-08-16 (Review 23)

Scope: 30 entries in 06-reference/ dated 2026-08-11 → 2026-08-15 (--since 7d --limit 30 --fix). Reviewed as 5 parallel batches of 6 (headless cron run).

DECISION-WORTHY FINDING — supersession rule produced a false-clean signal. The audit log shows [[2026-08-11-amazon-writing-style-tip-3]] failing I3/I8/I9 as of 2026-08-11T18:04:42 through 2026-08-12T00:00:32, then absent from the failure list from 2026-08-12T06:00:42 onward (multiple runs, all Fail: 0 or listing only unrelated files) — by the step-1.5 supersession rule as written, this reads as "fixed, drop from pre-failure set." But the reviewing batch read the file directly on 2026-08-16 and found author, ## Why this is in the vault, and ## Mapping against Ray Data Co were still genuinely missing — the file was never actually repaired. Either the audit script has a false-negative gap for this file (something about it is silently exiting its I8/I9 checks without a real pass), or the script isn't re-scanning files outside its rolling --since window on later runs and is carrying forward a stale "not in this run, so not failing" read. This is a real bug signal, not a rubric-fit question — worth a direct audit-newsletter-outputs.py run against this file (or its now-fixed state) to see what it reports today, since the fixed version should trivially pass and would confirm whether the script's checks work at all on this file.

improve_processed: 2026-08-17

/improve autonomous run — 2026-08-17

Diagnosis verification (guardrail step 3 + recurrence check 5/7)

Review 23 offered two hypotheses. Both were tested directly before any edit:

Census (guardrail check 6 — the review sampled, I censused)

Low-risk fixes applied: 1

  1. self-review/SKILL.md — step 1.5 scope-shortcut guardrail. The 2026-08-10 supersession rule's window condition was correct as written; Review 23 short-circuited it, building the pre-failure set from the two newest runs' Fail: 0 (both Since: 2026-08-14, 6 files audited) and declaring the set empty. Step 1.5 now states explicitly that a clean run clears only files inside its own window, forbids deriving the set from the newest run's fail count, and requires resolving each candidate against the most recent run whose Since: covers that file's date — falling back to a direct --since <file-date> re-run when in doubt. Worked example + the 55-failure census embedded.

Structural changes queued: 1

ESCALATION — both standing tasks still Blocked (live-verified in Notion this run)

No-ops

Process notes

2026-08-23 (Review 24)

improve_processed: 2026-08-24

/improve autonomous run — 2026-08-24

Finding 1 — I12: closed by founder, and the data agrees (recurrence check 1)

Finding 2 — the cross-link dings are a contract inversion, not per-file drift (guardrail step 3)

Review 24 filed its two cross-link B-grades as unrelated one-offs: "3 files hit exactly one soft criterion each, no pattern across them." Opening the files says otherwise.

Low-risk fixes applied: 2 (both in process-newsletter/reference/vault-note-schema.md)

  1. ## Related ≥2 threshold now explicitly counts vault files only. Removed the "or a memory slug" substitution licence, kept memory slugs welcome as supplementary and never-stripped, embedded the 29/369 census as the rationale. Closes the schema↔rubric inversion above.
  2. Named the I12 absorption failure mode + added the thought-leadership off-ramp for issues with genuinely no secondary items. Documentation only — no gate, per the founder's archived-task decision.

Structural changes queued: 0

Nothing in Review 24 is structural. Finding 1 is founder-closed and statistically resolved; queueing anything against it would re-open a decision the founder already made. Finding 2 was a wording fix to the authoring contract.

ESCALATION — four /improve proposals Blocked, live-verified via Notion SQL this run

None can self-clear; each needs one founder click. Ordered by age:

Worth noting against the last one: this run's I12 census had to be computed by direct filesystem scan precisely because the audit cron's rolling window cannot see the 41 pre-July failures. The blocked task is the reason the recurrence narrative stayed stale for two months — the cron could only ever show the newest failures, never the trend.

No-ops

Process notes

2026-08-30

/improve autonomous run — 2026-08-31

Finding — the B-grade outlier is founder-closed, and the census agrees

Low-risk fixes applied: 0 — 1 BLOCKED BY CLASSIFIER, needs one founder command

The one genuine improvement this cycle is a meta-fix to /improve itself, not to a downstream skill: amend Recurrence check 1 to name Archived as a third disposition (Done / Blocked / Archived). An archived task is a decision on record — report the pattern closed and census it, never re-queue or re-escalate it as live.

The edit was blocked by the auto-mode classifier (skill self-modification under ~/.claude/**). Per the classifier hard-gate rule, stopped after one attempt and routed here rather than working around it.

Structural changes queued: 0

Nothing in Review 25 is structural. Its single finding is founder-closed and statistically resolved.

ESCALATION — the same four /improve proposals still Blocked, live-verified via Notion SQL this run

None can self-clear; each needs one founder click. Cycle counts incremented; zero movement since 2026-08-24.

As last cycle, the why-in-vault census here had to be computed by direct filesystem scan because the audit cron's rolling window cannot see the 501 pre-May failures — the same blind spot 3bff7d49…668d exists to fix. Two consecutive cycles have now had their headline finding depend on a scan the cron structurally cannot perform.

No-ops

Process notes

2026-09-06 (Review 26)

/improve autonomous run — 2026-09-07

Finding 1 — copy-paste-wall vs attributed primary-source quotes: 6th cycle, now queued

Finding 2 — mtime sort is unsafe in this vault (LOW-RISK, APPLIED)

Finding 3 — three out-of-scope audit failures

Meta-fixes to /improve itself: 2 applied

ESCALATION — four /improve proposals still Blocked, zero movement for a 2nd straight cycle

Live Notion SQL reads this run, not review-log snapshots. None can self-clear; each needs one founder click.

No-ops

Process notes

2026-09-13 (Review 27)

/improve autonomous run — 2026-09-14

ESCALATION: five /improve proposals Blocked, no founder movement

Live Notion fetches this run. None had been edited since the dates below. Each needs one click.

2026-09-20

/improve autonomous run — 2026-09-21

2026-09-27

/improve autonomous run — 2026-09-28

ESCALATION: five /improve proposals still Blocked — 3rd cycle without founder movement

Verified live this run by SQL query against the Task Board data source (notion-fetch by raw UUID 404s for all five — a connector/workspace-scope quirk, not deletion; the SQL query returns them with current Status, so these figures are ground truth, not review-log snapshots). Day counts as of 2026-09-28. Each needs one click.

A Blocked task is open but cannot self-clear (guardrail check 1). These five have now absorbed a combined ~296 task-days of parking. The 2026-09-21 run did not surface them at all, which is itself the failure mode the check exists to prevent — silence reads as resolution.

2026-10-04

DECISION NEEDED: [[2026-09-29-josharosen-acp-emerging-interface]] filename doesn't match its source attribution (I10, audit-confirmed). Rename to a source-consistent filename, or accept the filename/source mismatch as-is? The file has 10 existing cross-links, so a rename isn't a self-review-safe edit — it needs a coordinated pass.

improve_processed: 2026-10-05

/improve autonomous run — 2026-10-05

The skip-stub flag, re-diagnosed

The audit-failed set, re-checked at source

Low-risk fixes applied: 4 files + 1 artifact

  1. self-review/SKILL.md — skip-stub bypass surfaced INLINE on all four rubric rows it governs, with the sponsor-specific reasoning and the 5-of-5 census on the Bias/sponsor row, plus an explicit "do not add the section under --fix." Also added content_type: synthesis to Frontmatter carve-out 3. Changelog entry added.
  2. audit-newsletter-outputs.py — synthesis added to the is_source_corpus content_type bypass (I3/I10). Verified: the 09-28 file's I3 cleared while the negative controls still fail. Header changelog entry added, carrying a note that this is the 9th content-class carve-out.
  3. process-youtube/SKILL.md — Mode 2 step 6 gained an explicit synthesis-note frontmatter contract (content_type: synthesis required and it is what the detector keys on; no newsletter_format; sponsored = union of the cluster with its own ## Sponsorship section when true — a synthesis note is not a skip-stub; Related wikilinks every cluster member). Step 6 previously specified only the filename and the body's job, which is why the frontmatter was improvised. Changelog entry added.
  4. rdco-doctor.py — three fixes, 47 violations → 39, all 8 removed were false positives. (a) C1: non-skill bucket dirs (_shared, any _*, synced) excluded from the registry rather than reported as missing SKILL.md — flagged as known false positives on 2026-09-21 and 2026-09-28 and never excluded, and they were also inflating C7 by two permanent "dark skills." Exclusions are PRINTED in the summary so the carve-out stays visible rather than silently swallowing a directory. (b) C6: repo scripts/ dirs under ~/Projects are now searched, restricted to clone roots (.git is a directory) so the 42 ephemeral sw-* worktrees cannot make the verdict depend on today's checked-out branches. (c) C6: .mjs/.js/.ts refs no longer require the executable bit — node scripts run via node/npm, never a shebang — while existence is still required. Header changelog entry added.
  5. [[2026-09-28-moonshots-live-awards-synthesis]] — ## Sponsorship section added; re-audited clean.

Structural changes queued: 0 — and that is a deliberate call, not an empty result

Five /improve proposals are already Blocked on founder judgment (below). Queuing a sixth would add load to a queue that has not moved in four cycles. Two findings that would otherwise be queue candidates are instead recorded as evidence for proposals already sitting there:

DECISION NEEDED (carried from the 2026-10-04 review, unresolved)

[[2026-09-29-josharosen-acp-emerging-interface]] — filename prefix is the author's handle ("josharosen") but source: X (article), so I10 fires. Confirmed live again this run. Rename to a source-consistent filename, or accept the mismatch and carve X-article notes out of I10? The file carries 10 cross-links, so a rename needs a coordinated link pass, not a self-review edit. Not actioned either way.

ESCALATION: five /improve proposals still Blocked — 4th consecutive cycle

Status live-verified this run by SQL query against the Task Board data source (notion-fetch by raw UUID still 404s; the SQL query returns current Status, so these are ground truth, not review-log snapshots). All five remain Blocked. "Blocked since" dates carry from the 2026-09-28 verification; day counts recomputed to 2026-10-05. Each needs one click.

Combined parking: ~331 task-days. A Blocked task is open but cannot self-clear (guardrail check 1).

No-ops

Meta-loop (/improve itself)

Added guardrail check 8: grep the target for the rule before you add it. A review that says "skill X failed to do Y" invites you to write a rule for Y; if the rule is already there, the defect is reachability, not absence, and the fix is placement — move it to the point of use, never leave one copy and add a second. Diagnose placement by contrast: the carve-outs the same review DID honour tell you where a rule has to live to be load-bearing. This is not covered by check 3 (write-path — whether the skill you edit is on the artifact's write path) or check 2 (scans recent cycles' Changelogs, so a contract written in May and never re-touched is invisible to it). Without this check, this run would have shipped a ## Sponsorship requirement onto skip-stubs that contradicts both the audit script and the rubric's own carve-out.

2026-10-11

DECISION NEEDED

  1. [[2026-10-07-shopify-simgym-shopgym-simpersona]] — audit-failed on literal section-header strings and a missing newsletter_format field, but the underlying content substantively satisfies the rubric (B/11). A 10-minute manual touch-up (rename 2 headers to the literal strings, add newsletter_format) would clear the audit without a rewrite. Worth doing, or leave as a known literalism gap between the deterministic audit and the semantic rubric?
  2. Five /improve proposals remain Blocked as of the last verified check (2026-10-05) — not re-verified this cycle. Still each needs one founder click (see 2026-10-05 entry for links).
  3. [[2026-09-29-josharosen-acp-emerging-interface]] I10 filename/source mismatch — still unactioned, not re-verified this cycle.

improve_processed: 2026-10-12

/improve autonomous run — 2026-10-12

The fix-threshold leak, censused

The I3 fall-throughs, deliberately not carved out

A full-year audit also surfaced that all three live I3 failures in Sep–Oct are one class: top-level 06-reference/ notes that are not newsletter issues, carry no content_type, and have a descriptive source — [[2026-10-04-kutay-gtm-in-git]] (X article), [[2026-10-04-starter-story-pat-walls-assessment]] (Ray-authored assessment) and [[2026-10-07-shopify-simgym-shopgym-simpersona]] (arXiv paper trio). audit-newsletter-outputs.py left an explicit tripwire on 2026-10-05: "this is the 9th content-class carve-out… adding a 10th is a signal to re-escalate that decision, not a victory." The tripwire has now fired and I did not act on it — adding carve-outs here would be the 10th, 11th and 12th classes. Recorded in the script header with the three named casualties so the next extender sees the evidence at the point of decision. Note the X-article variant of this proposal is Archived (guardrail check 1: a founder decision on record), which makes the Kutay flag a false positive by that decision and strengthens the case for inverting the rubric rather than extending it.

Low-risk fixes applied: 2 skills + 1 script + 1 artifact

  1. self-review/SKILL.md — new mechanical-defect fix path in step 4, running BEFORE the score gate and independent of it. Fixes header-literalism codes (I8/I9/I12) on any file in the audit pre-failure set regardless of score, by renaming a drifted header to the canonical string when functionally-equivalent content is already present. Explicitly EXCLUDES I3 (open founder decision — do not auto-add newsletter_format), I10 (needs a coordinated inbound-wikilink pass) and I11 (needs a judgment about sponsor-bias surface). Also corrected step 1.5's wiring note to name step 4 as a consumer. Changelog entry added.
  2. improve/SKILL.md — guardrail check 9 (see Meta-loop below). Changelog entry added.
  3. audit-newsletter-outputs.py — tripwire-tripped note in the header changelog, with the three named I3 fall-throughs and the reason no carve-out was added. Comment-only; verified the script still parses and runs (--since 2026-10-07, rc=0).
  4. [[2026-10-07-shopify-simgym-shopgym-simpersona]] — the first artifact fixed by the new rule, as its validation: ## Why in vault → ## Why this is in the vault, and ## What a tiny store can actually replicate → ## Mapping against Ray Data Co with the original specific title promoted to a bold lede so the framing survives the rename. Re-audited: I8 and I9 cleared, I3 still flagged — exactly the designed outcome, and the rule now carries this as its worked example.

Structural changes queued: 0 — deliberate, 2nd consecutive cycle

The one structural candidate this cycle (invert the frontmatter rubric so the newsletter contract is opt-in) is already Blocked and is the very proposal the tripwire points at. Queuing a sixth proposal onto a queue that has not moved in five cycles would add load, not signal (guardrail check 1: Blocked is open — do not re-queue a duplicate; escalate instead).

ESCALATION: five /improve proposals still Blocked — 5th consecutive cycle

Status live-verified this run by SQL query against the Task Board data source (returns current Status, so this is ground truth rather than a review-log snapshot). All five remain Blocked. Day counts recomputed to 2026-10-12.

Combined parking: ~366 task-days, up from ~331 last cycle. A Blocked task is open but cannot self-clear.

DECISION NEEDED (carried, unresolved)

[[2026-09-29-josharosen-acp-emerging-interface]] — I10 filename/source mismatch, carried from 2026-10-04 and 2026-10-05, now 3rd cycle. Rename to a source-consistent filename, or carve X-article notes out of I10? The file carries 10 cross-links, so a rename needs a coordinated link pass — which is exactly why the new step-4 path excludes I10 rather than auto-fixing it. Not actioned either way.

No-ops

Meta-loop (/improve itself)

Added guardrail check 9: a defect the review declined to fix is a permanent leak, not a closed item. Reviews report two kinds of defect — ones they fixed, and ones they found and left alone because their own gate said not to. The second kind reads as restraint, so it is never counted as a failure, and nothing re-opens it because the next cycle scores a new window. Census the threshold's own vocabulary; when the shape recurs, the defect is the gate, not the artifact, and the fix is a defect-class escape hatch beside the score gate. This is the mirror of check 4 (repair happens every cycle but the write-time gate is missing); here repair never happens at all. It is also outside check 1, because a within-threshold decision never generates a task to fetch. Without this check, this run would have answered DECISION NEEDED #1 with a 10-minute manual touch-up on one file and left the gate that produced it untouched for a seventh cycle.

2026-10-18

DECISION NEEDED

None new this cycle — both defects found were cleanly resolvable within the --fix mandate (score-gate fix + header-literalism rename) without a founder judgment call. The carried items above (5 Blocked /improve proposals, 1 carried I10 mismatch) remain the standing asks from prior cycles.

improve_processed: 2026-10-19

/improve autonomous run — 2026-10-19

The ad-hoc filing class: a conflicting contract, not a missing one

The real find: the proposal queue was shredding founder decisions unread

Pulling guardrail check 1 on the one task that matched this pattern turned up the mechanism behind ~6 months of stalled patterns.

Low-risk fixes applied: 2 files

  1. 02-sops/content-intake.md — reconciled with the canonical filing contract (the on-path fix for the ~9-cycle pattern). Step 5 now states the contract's required frontmatter (date, type as vault class, content_type as media shape, tags ≥2), the three verbatim body headings, and the sponsorship rule, with a pointer naming the contract as the floor that wins on conflict; kept published/source/author/shared_by as real provenance and noted shared_by (not shared_via) is the form in use. Step 3 no longer reads as permission to omit ## Related. Amendment log added.
  2. improve/SKILL.md — guardrail checks 10 and 11 added, check 1 amended inline (an Archived status is not self-evidently a decision; fetch the body first), and step 4 rewritten to queue proposals as Blocked + an HQ decision rail with a duplicate-check-including-Archived pass. Changelog entry added.

Structural changes queued: 1

ESCALATION: five /improve proposals still Blocked — 6th consecutive cycle

Live-verified by SQL against the Task Board this run (ground truth, not a review-log snapshot). All five remain Blocked. Day counts to 2026-10-19.

Combined parking: ~401 task-days, up from ~366. A Blocked task is open but cannot self-clear.

Correction to the 2026-10-12 run

That run treated the X-article proposal's Archived status as a founder decision and concluded the 2026-10-04-kutay-gtm-in-git I3 flag was "a false positive by that decision". The page shows no sign of founder engagement (blank body, never edited after creation), so that conclusion rested on a decision that does not appear to have been made. The I3 question is open, not settled — it now sits with the Blocked rubric-inversion proposal.

No-ops

Meta-loop (/improve itself)

Two new checks, both earned by this run.

Check 10 — an Archived proposal is only a decision if a human archived it, and your own queue is on a 14-day timer. This is the most expensive bug this skill has had: it made the skill's own escalation channel self-clearing, then taught every later run to treat the silence as consent. The tell was available for months — "Auto-archived: no activity for 70 days" is sitting in the page body — but check 1 said to read Status and stop. Status is a summary; the body is the evidence.

Check 11 — "no skill writes this" is not "no instruction governs this." Check 3 asks which skill authors a class and, finding none, concludes only a post-write gate can reach it. But ad-hoc sessions read SOPs, so a SOP is a write-path document. Nine cycles applied check 3 correctly in letter and still missed the fix, because none of them looked outside ~/.claude/skills/. The generalization: when a path looks ungoverned, the question is not "which skill writes this" but "which document does the writer read" — and a wrong instruction at the point of use is both more likely and cheaper to fix than a missing gate.

2026-10-25

DECISION NEEDED

  1. [[2026-10-20-winning-by-design-spiced-framework]] I3 looks like a rubric/audit false positive, not a vault-quality defect. content_type: vendor-page, source: "Winning by Design (company website)" — there is no newsletter sender, issue, or digest structure here, so newsletter_format doesn't conceptually apply. The audit's I3 check and this skill's frontmatter carve-out list don't yet exempt content_type: vendor-page (or company-website sources generally) the way they exempt YouTube/research-brief/concept/source-corpus classes. Recommend extending the carve-out list rather than leaving this as a recurring false flag. Did not change the audit script or this skill file myself — that's an /improve-shaped edit, not a self-review content fix.
  2. The 2026-10-19 ad-hoc-filing SOP fix may not be taking effect at write time. See systemic pattern above — both of this cycle's genuine C/D defects post-date the fix by 3 days and show the same symptoms the fix was meant to prevent (non-canonical headers, missing content_type, one missing ## Related entirely). Worth checking whether whatever wrote these two files (a session, a different skill, a manual Write) actually reads 02-sops/content-intake.md before filing, or whether the amendment needs to propagate somewhere else (e.g. a skill prompt that currently duplicates the old contract instead of pointing at the SOP). This is the kind of trend that should go back to /improve rather than get re-diagnosed identically a 10th time.

improve_processed: 2026-10-26

/improve autonomous run — 2026-10-26

DECISION NEEDED item 1 — vendor-page I3 carve-out: DECLINED, escalated instead

The review recommended extending the I3 carve-out list to content_type: vendor-page so 2026-10-20-winning-by-design-spiced-framework.md stops false-flagging. Its diagnosis is right (a company sales page has no sender and no issue, so newsletter_format is a category error on it) and the fix is wrong, because it would be the 11th carve-out and guardrail check 2 forbids enumerating against an unbounded space. Censused before deciding:

Re-escalated on the existing Blocked proposal 3aaf7d49…c5d7 ("invert the self-review frontmatter rubric"), 91 days blocked, 7th consecutive cycle, census appended to its page body and Priority raised Medium → High so /check-board surfaces it (reversible; the decision itself remains the founder's). No duplicate task created — check 1.

Correction to the record (guardrail check 12, new this run). Check 10 landed on 2026-10-19 and proved that Archived on an /improve proposal is a /check-board 14-day timeout, not a founder judgment. But the 2026-10-12 run had already written that phantom decision into two surfaces that actually execute — self-review/SKILL.md step 4 ("the adjacent proposal … was Archived — a decision on record that X-article notes should NOT be forced into the newsletter contract") and the audit-newsletter-outputs.py header ("which makes the Kutay flag a false positive by that decision") — and the 10-19 run amended the guardrail without sweeping them. So for a week the skill set was still reasoning from a decision nobody made. Verified today by fetching 34cf7d49…ac30: blank page, page_last_edited_at identical to Created (2026-04-24T08:10:39.186Z), never touched. Both citations corrected in place; the I3 exclusion itself still stands, now resting on the Blocked proposal alone. Changelog/review-log history left intact as audit trail.

DECISION NEEDED item 2 — ad-hoc filing SOP "not taking effect": re-diagnosed, fixed on-path

The review is correct that the 2026-10-19 02-sops/content-intake.md amendment has not changed behaviour: both 2026-10-22 defects post-date it by 3 days with the same symptoms. It asked whether the amendment needs to propagate somewhere else. It does, and the gap is reachability, not content — the SOP as amended is correct (step 3's licence-to-omit-## Related reading is closed, step 5 now defers to the canonical contract).

grep -rl "content-intake" ~/AGENTS.md ~/SOUL.md ~/.claude/skills/ ~/.claude/state/working-context.md returns nothing (one hit in improve/SKILL.md, its own changelog). Nothing in always-loaded session context names the SOP, so an ad-hoc founder-dialogue session filing a shared link has no path to it. A correct document that no session loads is not a fix — which is why ~10 cycles of correctly locating the write path never moved the number.

Fixed with the cheapest rail that is actually always in context: a new auto-memory entry, feedback_founder_shared_filing_contract.md, plus its MEMORY.md pointer. MEMORY.md loads every session by construction, the write is low-risk and reversible, and it needs no founder decision — unlike editing ~/AGENTS.md (founder-owned) or installing a PostToolUse hook (harness-level, structural). The memory states the frontmatter floor, the three exact-string headings, the post-write audit-newsletter-outputs.py --since <date> check, and the standing exception that a bare I3 on ad-hoc content is expected and must NOT be "fixed" by inventing a newsletter_format value while 3aaf7d49…c5d7 is open. Guardrail check 11 amended to require this second half (instruction exists AND something loaded names it) before any SOP reconciliation is recorded as a fix.

Low-risk fixes in detail

  1. self-review/SKILL.md step 4 — added I2 to the mechanical-defect fix path, scoped to the unquoted-colon case only (quote-wrap the scalar, verify by yaml.safe_load round-trip + re-audit); other I2 causes (unclosed brackets, indentation, tabs, duplicate keys) explicitly stay off it. The review flagged this itself after fixing 2026-10-21-moonshots-chamath-… off-list. Censused first per check 6: the shape has fired on live files three times — Review 10 (2026-05-05, books trio, queued as a manual sweep that never happened), 2026-10-05 (every-sam-altman-dots, author: carrying interview: Dan Shipper…), 2026-10-25 (moonshots-chamath, author: carrying Guest: Chamath…). In both October cases the fixing run did the fix and then noted it was off-list, so the list was the defect.
  2. self-review/SKILL.md step 4 — I3 exclusion rationale corrected (check 12 sweep, above).
  3. audit-newsletter-outputs.py header — 2026-10-26 entry recording the census, the refusal to add carve-out #11, and a partial retraction of the 2026-10-12 note's phantom-decision claim. Script syntax verified (ast.parse) and re-run post-edit.
  4. Auto-memory feedback_founder_shared_filing_contract.md + MEMORY.md pointer (item 2, above).

Standing escalation — 6 Blocked /improve proposals, 446 task-days combined

Not re-queued (check 1), cannot self-clear, so listed every cycle until a click lands:

Task Blocked since Days
389f7d49…c62e 06-reference folder restructure 2026-06-24 124
39cf7d49…624e self-review scope for 08-tooling 2026-07-13 105
3aaf7d49…c5d7 invert the frontmatter rubric 2026-07-27 91
3bff7d49…668d audit cron 1-day window 2026-08-17 70
3d4f7d49…f02b copy-paste-wall misfire 2026-09-07 49
3fef7d49…26f8 stop auto-archiving proposals unread 2026-10-19 7

3aaf7d49…c5d7 is the one with live, rising evidence behind it (231 I3 violations, 7 in 22 days) and the one whose absence forces every other run to keep declining the same carve-out.

Not actioned, noted for next cycle

2026-10-26-lieberman-head-of-ai-5b-real-estate-notes.md (filed today, outside the reviewed window) already carries I3 + I6 + I10 — including sponsored: 'unknown' as a string, which is an I6 type break not seen in recent cycles. It is the 7th October I3 instance and the first test of whether the new memory rail changes write-time behaviour. Next cycle should check it specifically rather than scoring a fresh window blind.

2026-11-01

DECISION NEEDED

improve_processed: 2026-11-02

/improve autonomous run — 2026-11-02

The finding: /self-review has been scoring 30 of ~46 files a week for 18 weeks

The review's own systemic section said "continuing evidence (not new)" and pointed at the standing Blocked rubric-inversion proposal. That read is correct as far as it goes, and I verified it (check 1): 3aaf7d49…c5d7 is still Blocked, Priority High, 98 days (the review said 97; recomputed from 2026-07-27). No duplicate queued.

But the real finding this cycle was in the review's scope line, not its systemic section. It mentioned in a parenthetical that "the eligible window actually held 39 files, 9 oldest excluded by the 30-cap", and separately flagged 2026-10-26-anthropic-s1-teardown.md as I3-failing and "outside this review's 30-file scored window" — two adjacent bullets that were never connected.

Censused over 18 consecutive weekly windows (top-level 06-reference/, filename dates):

Windows exceeding the --limit 30 cap 18 of 18
Eligible files per window 37–59 (mean ~46)
Dropped per window 7–30 (mean ~16)
Total dropped ~285 files

Because the cron runs --since 7d, next week's window has moved on: a file pushed out by the cap is not backlogged, it is unreachable. And sort -r | head -30 sacrifices the same oldest end of the window every time — exactly where already-filed, never-repaired defects sit. This is recurrence-guardrail check 9's shape (a defect nothing re-opens) living in the candidate list itself, and it is also the mechanism behind the long tail of I3 violations earlier cycles kept re-encountering: a 2026-11-01 census of the full vault is not needed to see that 231 live I3s and a 16-file/week blind spot are the same story.

Verified against the primary sources before acting, per step 3's diagnosis guardrail: the cron line is 0 7 * * 0 | system | sonnet | - | /self-review --since 7d --limit 30 --fix in scheduled-jobs.txt:69; step 1's own snippet is sort -r | head -30; and grep -niE "backlog|ledger|carry|catch.?up|unscored|already scored" over self-review/SKILL.md returns nothing, so no catch-up mechanism exists and this is a genuine absence, not the misplacement check 8 warns about.

Applied (low-risk, no decision needed): step 1 now requires computing the real denominator, appending cap-excluded filenames to an overflow ledger at ~/.claude/state/self-review-unscored.md, and spending 10 of its 30 slots on the oldest unscored backlog — per-run cost unchanged at --limit files, so this is a selection-policy change, not a budget change. Step 3 now requires a Coverage: N scored / M eligible / K dropped line, because Entries reviewed: 30 is phrased as a result rather than a fraction and that phrasing is how this hid for 18 weeks. Ledger seeded with a 130-entry reconstruction from the last 8 windows, labelled as reconstructed (per-run scoring records were never kept, so a redundant re-score is possible — cheap, versus a permanent miss).

Queued (structural, founder's call): carry-forward bounds the loss but cannot clear it — intake ~46/week against a 30/week budget means the backlog still grows ~16/week. Raise the limit (more subagent spend), keep 30 and accept now-documented sampling, split to fortnightly, or tighten eligibility. Ray's read in the task: keep 30, since the drop is now recorded and rotated rather than silent. Scope-checked distinct from 3bff7d49…668d (the audit script's 1-day cron window — different tool, different mechanism) and from 39cf7d49…624e (which directories are eligible, not how many files are sampled).

Second fix: the scorer's phantom defects now have to be verified by the producer

The review's subagent-fan-out QA note (a 3-way fan-out independently reporting a missing frontmatter type: on two files that both carry type: reference) is n=1 for that exact vocabulary, which on its own is below the restraint threshold. Censused the broader class instead — the scorer reported a defect that direct file inspection falsified — and it has fired three cycles by three different routes: 2026-07-13 (missing newsletter_format misattributed to /process-newsletter; 4 of 5 files were YouTube notes correctly omitting it), 2026-10-04 (docked a skip-stub's sponsor point against a carve-out self-review/SKILL.md itself grants), 2026-11-01 (the phantom type:, which cost 2026-10-28-moonshots-… a full grade — reported 11/13 B, true 13/13 A).

Three cycles clears the threshold. self-review/SKILL.md has no fan-out dispatch contract at all (the batch split is an ad-hoc parent decision each run — "3 batches of 10", "5 subagents of 6"), so there was no rule to relocate. Added one to step 3 at the point of use: an absence claim about a frontmatter field or canonical heading is mechanically checkable, so it must be checked against the raw file before it becomes a score, a grade, or a systemic pattern. Note the asymmetry this closes — /improve's own step-3 diagnosis-verification guardrail protected the consumer of these reports since 2026-07-13, while the producer had nothing. The 2026-11-01 run caught its own error only because it chose to verify.

Meta-loop: guardrail check 13

Per the "no one-off work" test. None of checks 1–12 would have surfaced this run's finding, and I checked each: no task to fetch (1), no prior wording fix (2), correct write path (3, 11), artifacts never auto-fixed (4), no gate credited (5), every reported instance real (6), no logged failure involved (7), rule genuinely absent not misplaced (8), nothing declined on a threshold (9), no status misread (10, 12). Check 6 censuses the instances a review reports; nothing censused the set it drew from. Added check 13 — a review's scope line is a coverage claim, so census the denominator, with the decisive follow-up question: can a dropped item ever come back? If the next window has moved on, it is a permanent leak, not a backlog. Generalizes past review logs to any report carrying a --limit, head -N, page_size or "top N" in its provenance.

Standing escalation — 6 Blocked /improve proposals, 488 task-days combined

Not re-queued (check 1), cannot self-clear, so listed every cycle until a click lands. All six statuses re-verified live this run, not taken from last cycle's snapshot:

Task Blocked since Days
389f7d49…c62e 06-reference folder restructure 2026-06-24 131
39cf7d49…624e self-review scope for 08-tooling 2026-07-13 112
3aaf7d49…c5d7 invert the frontmatter rubric (High) 2026-07-27 98
3bff7d49…668d audit cron 1-day window 2026-08-17 77
3d4f7d49…f02b copy-paste-wall misfire 2026-09-07 56
3fef7d49…26f8 stop auto-archiving proposals unread (High) 2026-10-19 14
40cf7d49…3513 self-review drops ~16 files/week (High, NEW) 2026-11-02 0

Eight other /improve proposals sit Archived; per check 10 those are /check-board 14-day timeouts, not decisions, and remain re-queueable.

Rail gap — flagged, not worked around

Step 4 requires an hq.raydata.co/decisions/ click-back page alongside the Notion task. I did not create one: ~/rdco-hq is on feature branch feat/async-cadence-parser-routing with uncommitted changes, so writing and prod-deploying a page from an unattended cron run would mean deploying prod off-main (feedback_prod_deploy_only_from_main.md). Task queued Blocked anyway rather than downgraded to To Do to compensate, and the gap is named in its Notes. Next interactive session: add the 4-option rail for 40cf7d49…3513. Also checked while there — the two older /improve rails (2026-08-17-audit-window-coverage-scope, 2026-08-30-x-write-rail-oauth-unblock) ARE committed on origin/main and return HTTP 302 (Cloudflare Access gate), so they are live; the untracked copies in the working tree are stale duplicates, not undeployed pages.

2026-11-08

DECISION NEEDED

/improve autonomous run — 2026-11-09

The review reported "Systemic issues: none found" — that was true and not the whole story

The 2026-11-08 review is a clean, well-run cycle: 30/30 A, 13.0/13, one verified mechanical I12 fix, honest Coverage: accounting (30 scored / 47 eligible / 27 dropped, ledger 147), and a correct re-audit to confirm the fix took. Nothing in it is wrong. But a 13.0/13, no systemic issues result is also a measurement of the reviewer, and nothing recorded it as one.

Censused (check 6 discipline, applied to the grade line rather than an instance count): the semantic rubric has returned ~13/13 and ~100% A for 10 consecutive cycles (2026-09-06 → 2026-11-08) and found no new systemic pattern in 7 of them. Over the same span the cheap deterministic gate kept finding real violations (14 files failing vault-wide; 2 of this cycle's 30), and this cycle's own double-signal: 0 line says the two signals no longer agree. So the expensive half of /self-review (30 subagent scorings/week) is confirming a contract the authoring templates already satisfy, while the grep finds everything actually broken. Never named in any prior cycle — grep -niE "saturat|ceiling|discriminat" over the full log returns only the 300-line conciseness ceiling.

Applied the low-risk slice (new check 14): /self-review step 3 now emits a mandatory Signal yield: {N} new / {M} consecutive none line, and must raise 0 new / 3+ consecutive under DECISION NEEDED as a question about remaining discriminating power rather than re-reporting it as "vault quality stable." Re-pointing or retiring the 30-scorings/week budget is structural; that went to the existing proposal as evidence, not to a new task.

Check 14 (new): measure your own escape hatch before classifying into it

Checks 1–13 all police how a review's input claims are read. None asks whether step 2's output surface works. It does not: 7 proposals Blocked at 7–138 days (median 84), none ever answered. Since 2026-06-24, classifying a pattern as "structural" has been functionally identical to deciding never to fix it — which means the low-risk/structural boundary has been quietly setting severity, not risk.

Two corrections fell out of it:

Audit-script signals (step 0)

DECISION NEEDED

Nothing new. But the standing queue is now escalated rather than noted, per check 1 and the rewritten step 6:

Age Proposal
138d 389f7d49…c62e 06-reference/ per-sender subfolders
119d 39cf7d49…624e self-review scope for 08-tooling/
105d 3aaf7d49…c5d7 invert the self-review frontmatter rubric (I3)
84d 3bff7d49…668d audit cron's 1-day window
63d 3d4f7d49…f02b copy-paste-wall misfire on cited primary sources
21d 3fef7d49…26f8 stop the proposal queue auto-archiving decisions unread
7d 40cf7d49…3513 --limit coverage gap (new evidence added this run — now a clearer call)

All 7 have an HQ decision rail. All 7 are unanswered. The cheapest unblock is 40cf7d49…3513: the new saturation evidence makes option (b) — keep the 30-file cap, accept documented rotation — the clear answer, and it costs nothing to confirm.

2026-11-15

/improve autonomous run — 2026-11-16

1. The review's 13.0/13 covered 3 files, not 30

The review's Scoring method line says "mechanical pass on all 30 + spot semantic read of 3. Not a full semantic read of all 30." The next lines still say Average 13.0/13 (A:30) and Signal yield: 0 new / 8+ consecutive. So 27 files got full marks on semantic criteria nobody read. This is also the first cycle where the run decided by itself to cut the 30-scorings/week budget, which is the exact question parked on 40cf7d49. It did disclose this, so it is not hidden. The problem is that the headline numbers and the streak still read as full coverage, and the streak is the evidence for that parked decision. Fix applied to self-review step 3 (the report template) and step 5 (the review-log template). Reports now split semantic-scored files from mechanical-only files and never credit a criterion that was not read. A partial-read cycle reports streak paused and no longer extends the "none found" streak. Choosing to cut semantic scoring has to go under DECISION NEEDED. The step-5 log template was also missing the Coverage, Signal yield and DECISION NEEDED lines that step 3 already required, so this run's DECISION NEEDED appears only in the ops log and not in this log. Added them. Checking last week's fix: the ops-log copy of this review raised 0 new / 8+ under DECISION NEEDED correctly, so the 11-09 signal-yield rule works.

2. The founder-shared filing failed I3/I11 because it followed the SOP (checks 11/12)

I opened the file directly (diagnosis check). 2026-11-13-building-always-on-agents.md uses source_url: (no source:) and ## ⚠️ Structural Bias (no ## ⚠️ Sponsorship). content-intake.md explicitly allows both: "source / source_url" and "## ⚠️ Sponsorship (or ## ⚠️ Structural Bias)". The audit rejects both (I3 requires source; H_SPONSOR matches only Sponsorship). Count across the vault: 2,605 notes use source: and 6 use source_url alone; 914 use the Sponsorship header and 6 use Structural Bias. The majority form is what the audit expects, so I changed the SOP rather than the audit. The invariants SOP's own line 57 also allowed Structural Bias, contradicting its own I11 row, so I corrected that too. Reachability (check 11 amendment): MEMORY.md → feedback_founder_shared_filing_contract → content-intake.md, so the corrected SOP will actually be read. The newsletter_format half of this file's I3 failure stays parked on Blocked 3aaf7d49. The 6 existing source_url-only files and 2 sponsored Structural-Bias files were NOT edited. They are known, the count is small, and the I3 half is parked anyway.

3. The 2026-11-10 changelog migration was incomplete (found via step 0)

13 skills still had ## Changelog inside SKILL.md, including building-an-exo at 7.6 KB and remix. self-review also had 11 history entries stuck under ## Related, where migrate cannot see them, which left SKILL.md at 42 KB. I ran migrate --all (dry-run now reports 0) and moved the 11 stray entries by hand. self-review SKILL.md went from 42.4 KB to about 31 KB.

Audit signals (step 0)

Not actioned (check 9)

The vault-wide audit-failing count went from 14 to 15 (I3 x10, I11 x3, I10 x2). I3 is parked on 3aaf7d49. Fix #2 should stop new I11 failures on the founder-shared path. The other I11/I10 files have not been diagnosed this cycle. Next run: list them and check whether they share a write path.

DECISION NEEDED

Nothing new. 7 Blocked proposals, all unanswered. The quickest one is still 40cf7d49 (self-review budget). The run has now started making that call itself, so a yes/no answer would turn an unofficial practice into a decided one.