Ops

active ray + harness

Building the COO agent and the harness it runs in. Vault, skills, MCP, observability, deployment rails.

hq.raydata.co

at a glance

Skills
65+
Cron loops
13
Vault docs
2913

open decisions 6

open medium
Audit cron's 1-day window — widen scope to catch the 55 invisible failures?
The nightly audit-newsletter-outputs.py cron's ~1-day --since window permanently drops unrepaired failures from scope after 24h; a direct --since 2026-06-01 run shows 55 live failures invisible to the daily cron. Self-review guardrail already shipped (stops false-clean reporting) but the coverage hole itself needs a founder call: (a) weekly wide-window (~60d) re-audit cron, or (b) carry-forward re-inclusion of unresolved failures; plus whether --fix should ever be wired to this (Ray recommends no, read-only surfacing for now).
2026-08-17
open medium
Self-review rubric inversion — scope this or park it?
Notion task 3aaf7d49-36d1-8186-be95-f0562b79c5d7 has no scope beyond its title. Traces to a real /improve finding (2026-07-27, from review-log.md Review 20 cycle): the newsletter-frontmatter carve-out list has been extended 9 times in 3 months; proposed fix is a single provenance: field both graders branch on. No build spec exists yet.
2026-07-28
open medium
Vault 06-reference restructure — grant the Bash allow rule?
06-reference/ has grown to 2,317 files, 4.6x the threshold that triggered the per-sender-subfolder proposal. Blocked since 2026-06-26 on an auto-mode classifier denial for SKILL.md-edit subagent dispatch — needs a founder-granted Bash allow rule under ~/.claude/skills/, which would also unblock the related deep-fetch-threshold proposal.
2026-07-26
open medium
/behavior-critic — scope + activation
New source-BLIND critic built that exercises artifacts (runs scripts, clicks forms, calls endpoints) against a written behavior contract, closing the gap named in the openclaw/autoreview teardown. Tested against 6 fixtures: cli/api (3 flawed + 1 clean) all correct with zero false approves; web-interactive blocked by a Playwright browser-lock contention issue, discipline held (no fabricated verdict) but that class unproven end-to-end. Dormant pending founder call on scope priority, web-interactive retest, and whether/how to wire a gate.
2026-07-26
open high
/supervise pattern — activate on the first write-path?
Fresh-eyes supervisor sub-agent built and tested (9 scenarios, 7/7 flawed actions caught, zero false approves), modeled on Project Vend Phase 2 CEO-oversight. Dormant pending founder call on action-class tiering, settings.json hook splice for Gmail create_draft, and future coverage extension.
2026-07-25
open high
verify-action hook splice — activate in settings.json?
Build complete, 6/6 fixture tests pass (em-dash, iMessage chat_id format, Discord external @mention, iMessage length rules). Only remaining step is splicing the PreToolUse hook into settings.json to make it live. Page existed on disk since 2026-06-24 but was never registered in the decisions index, so it never reached the founder.
2026-06-24
closed decisions 14
archived
/improve proposal: deterministic I12 heading validation in process-newsletter WATCH path
→ ARCHIVE
2026-06-30
decided
/weekly-bet-review skill — unblock or archive?
→ ARCHIVE
2026-05-23
decided
Alpaca brokerage account — was yesterday exploration only, or account opened?
→ approve
2026-05-12
decided
Multi-agent skill build-out pipeline with spec gating — file as Notion ticket?
→ APPROVE with refinement — meta pattern confirmed. Founder flagged 4 implementation questions (namespacing, triggering, per-domain test/critic shape, possible task split). Follow-up decision page spawned 2026-05-12-multi-agent-pipeline-implementation-shape. /build-website-discovery (the first worked example) stays Blocked until the shape decision lands.
2026-05-12
decided
hq.raydata.co PR #3 — merge to ship Phase 3.5 single-login UX
→ ARCHIVE — PR #3 was already merged. Parent 'Build hq.raydata.co v1' ticket stays In Progress for ongoing HQ work.
2026-05-11
decided
HQ Phase B: per-bet stack-view UI (Squarely first)
→ ARCHIVE — already merged. HQ is organized as desired; no more Phase B changes anticipated.
2026-05-11
decided
Stripe Atlas Partner Perks audit (5-min check in portal)
→ ARCHIVE — Cloudflare credits unavailable. AWS $5k + Google Ads $500 are the only valuable perks. Not immediate action; logged for future reference at 06-reference/2026-05-12-stripe-atlas-perks-audit-result.
2026-05-11
decided
Foldering decision: where do internal-state-snapshot docs live?
→ APPROVE Option B — state-reviews live at 01-projects/<bet>/state-reviews/. Squarely state-review migrated; SOP filed; audit-script interaction documented (no script change needed).
2026-05-11
decided
/email-extend skill: greenlight draft-and-watch email-loop pattern
→ APPROVE — build greenlit with closed-loop constraint: drafts/replies must route through /decisions/, /board/, and vault. No opaque side channels.
2026-05-11
decided
MOOC + foundational knowledge shortlist — commit to a path?
→ APPROVE — curriculum greenlit, D3 (Tooling U-SME, only paid resource) excluded. All 13 free/open resources approved as 6-12mo path.
2026-05-11
decided
Digital-manufacturing YouTube shortlist — pick channels to register
→ APPROVE — top-3 picks (CNC Kitchen, Teaching Tech, Product Design Online) registered to youtube-channels.json as tier-2 watch-only.
2026-05-11
approved
/weekly-plan + /weekly-status skill pair — greenlight build
→ APPROVE
2026-05-11
archived
Scaffold MAC + Sanity Check bet-stack YAMLs from audits
→ ARCHIVE
2026-05-11
decided
Click-back rail test (sms:ray@raydata.co)
→ Rail confirmed working on both iPhone + Mac. Pattern locked.
2026-05-09

recent vault notes 0

No vault notes match this bet's path prefixes (01-projects/process-newsletter, 01-projects/skill-improvements, 01-projects/self-review, 01-projects/hq-raydata-co, 02-sops, 08-tooling).