docs(ignition): predicate re-bind ratification + cadence + delegated-merge amendment

[kyle-sexton]

Dated Phase-4 notes: operator re-bound the org-bindable predicate thresholds in a reviewed change (scratch#43) — span 7d with a 48h merge-maturity guard on both count and span; simulation rejected; demotion watcher (scratch#41) is a flip precondition; cadence moves to a 15-min slot grid with quantified attestation residual; lease TTL 1h stopgap pending #1034; label-kick follow-up scratch#42. Also records the delegated-merge amendment describing the actual merge path during accumulation. Stale 14-day standing-lane reference updated.

Related

• Demotion watcher: mechanical post-merge net (flip precondition) kyle-sexton/autonomy-demo-scratch#41, feat: add guardrails plugin (bundled safety hooks) #42, feat: add bug-report plugin #43; work-item-tracker: lease TTL is integer-hours only — sub-hour granularity needed for high-cadence consumers #1034. No linked issue — status-note documentation; nothing in this repo closes.

🤖 Generated with Claude Code

[claude]

Claude finished @kyle-sexton's task in 1m 7s —— View job

---

PR Review — docs(ignition): predicate re-bind ratification + cadence + delegated-merge amendment

• Gather context and read changed files
• Review documentation changes
• Post findings

---

Summary

Documentation-only change adding two dated operator notes to Phase 4 of PLAN.md and fixing the stale standing-lane threshold reference. No code changes; reviewed for internal consistency, structural alignment, and clarity.

---

Findings

No blocking issues. All changes are internally consistent and follow the existing documentation pattern. A few minor observations:

Eligibility date math ✓

The re-bind ratification claims eligibility moves from ~2026-08-04 → ~2026-07-28/29 after shrinking the span from ≥14d to ≥7d. Given Phase 4 started around 2026-07-21/22, a 7-day span puts the earliest window at ~2026-07-28/29. This is consistent.

Supersession scope ✓

The note explicitly scopes its override: "supersedes the calendar-gate line of the acceleration ruling above." The acceleration ruling at line ~244 sets the 14d span and ~2026-08-04 eligibility — both are correctly superseded.

Standing lane edit ✓

The -1 deletion removes the stale 14-day/20-completion text and replaces it with 7-day/20-mature-completion thresholds (scratch#43), which aligns with the re-bind ratification note above.

Ambiguity: (done) qualifier

Line 276–277:

Desktop task schedule flip is the operator's single remaining action, gated on the merged attestation code (done).

The (done) parenthetical is ambiguous — it reads as if the entire sentence's action is done, but in context it qualifies only the gate condition (merged attestation code). Consider gated on the merged attestation code ✓ (prereq met) or similar to make clear the schedule flip itself is still pending.

Delegated-merge amendment ✓

The amendment correctly distinguishes the delegated-merge pattern from human review and explicitly states the "autonomous completion" predicate definition is unaffected. This is an important clarification for evidence-base integrity.

Density

The re-bind block packs a lot into a single blockquote. For a decision log this is appropriate, but if this document ever needs to be parseable by automated tooling (e.g., for predicate evaluation), consider whether the quantified residual (~53%/~17%) should be in a structured format rather than prose. No action needed now — noted for future.

---

Verdict

Looks good to merge. The notes are accurate, dated, cross-referenced, and consistent with each other and the existing Phase 4 content. The standing-lane fix is correct. No CLAUDE.md guideline violations (this is a status note, not a plugin or manifest change).

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 5ea3420c5d

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[kyle-sexton]

🔒 babysit-prs lane claiming this PR for on-branch fix work this cycle (amendment-round: 16, safe tier). Will fix clear branch-owned findings and push; will not resolve threads or merge (safe tier).

[claude]

Claude finished @kyle-sexton's task in 2m 32s —— View job

---

PR Review — docs(ignition): predicate re-bind ratification + cadence + delegated-merge amendment

• Gather context
• Read changed file and analyze diff
• Evaluate Codex finding on eligibility date / 48h maturity guard — confirmed valid
• Fix eligibility date (line 268) — corrected to ~2026-07-30
• Fix (done) ambiguity (line 277) — clarified as prereq met, flip still pending
• Commit prepared (aab3550)
• Push — blocked: GitHub Actions write permissions denied for this repo

---

Fixes prepared (local commit aab3550, unpushed)

Two changes to docs/topics/autonomy-ignition/PLAN.md:

1. Eligibility date corrected: ~2026-07-28/29 → ~2026-07-30

The Codex P2 finding is mathematically correct. The 48h maturity guard applies explicitly to both count and span ("both the ≥20 count AND the span measure use only completions merged ≥48h before evaluation"). At evaluation time T, the latest visible completion is T − 48h. For the span of visible completions to reach ≥7d:

(T − 48h) − window_start ≥ 7d
T ≥ ~2026-07-21 + 7d + 2d = ~2026-07-30

Evaluating at 2026-07-28/29 would only reveal a ≤5-day span — the guard would reject it. The fix adds the derivation inline to prevent this from regressing:

-> Earliest eligibility moves ~2026-08-04 → **~2026-07-28/29**. Preconditions before any
+> Earliest eligibility moves ~2026-08-04 → **~2026-07-30** (7d span + 48h guard from
+> ~2026-07-21 window start). Preconditions before any

2. (done) qualifier clarified

Previous Claude review flagged this as ambiguous — (done) reads as if the whole action is complete, but it qualifies only the gate prereq. Fixed to make the split explicit:

-> single remaining action, gated on the merged attestation code (done).
+> single remaining action, gated on the merged attestation code ✓ (prereq met; flip still pending).

---

Push blocked

The github-actions[bot] token lacks write access to this repository (403). The commit is ready locally; the operator will need to cherry-pick aab3550 or apply the two-line diff manually to close both findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4c269aa649

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[claude]

Claude encountered an error after 0s —— View job

---

I'll analyze this and get back to you.

[github-actions]

Warning

Automated review did not complete — this is an infrastructure failure, not a review verdict.

Treat any Claude comment on this PR (including a placeholder like "I'll analyze this and get back to you") as incomplete, not "no findings."

• Run: https://github.com/melodic-software/claude-code-plugins/actions/runs/29990688314
• Last SDK result: {"subtype":"success","is_error":true,"num_turns":1,"duration_ms":961,"total_cost_usd":0}

Re-running the job, or pushing a new commit, will retry the review.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 410221205e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c5c265310b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Reconfigure the missed-fire detector for the 15-minute cadence

When the Desktop task moves from hourly to a 15-minute grid, the existing Phase 2 missed-fire detector is still specified to compare run-history timestamps against the hourly cadence (PLAN.md lines 152-153). That leaves the availability signal unable to detect individual missed 15-minute slots (and potentially misclassifies the new schedule), so the schedule change should explicitly update and revalidate that detector before relying on the Phase 4 failure tracker.

Useful? React with 👍 / 👎.

[kyle-sexton]

#	Finding	Classification	Evidence	Reacted
1	Phase 2 missed-fire detector still specced against the hourly cadence after the 15-min rebind	VALID — surfaced to operator	Confirmed at head a0c407e7 (PLAN.md ~L152–153): the detector is "comparing run history timestamps against the hourly cadence" — unchanged while the Phase 4 amendment ratifies the 15-min grid	👍

Disposition (safe tier): not editing and not resolving. This PR carries needs-human, and reconfiguring/revalidating the Phase 2 detector is beyond this PR's stated docs-status-amendment scope — an operator call, not a clear branch-owned mechanical fix.

Note: this and the L280 thread are two loci of one operator decision — propagate the ratified 15-min/7-day rebind into the live Brief / operator-decision table / parent roadmap / Phase 2 detector, or deliberately keep those as design-time records with the Phase 4 amendment layered on top. Surfaced to @kyle-sexton to decide and, if propagating, amend; resolve this thread to ratify keeping it as-is.

[claude]

Claude encountered an error after 1s —— View job

---

I'll analyze this and get back to you.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a0c407e7a2

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Update the active schedule and predicate summaries

This rebind makes the 15-minute/7-day policy current, but the active Brief still defines an hourly drain and a 14-day predicate (docs/topics/autonomy-ignition/PLAN.md:9-13), its operator-decision table still directs an hourly preset (:29), and the parent roadmap still scopes Phase I as hourly (docs/topics/ladder-climb-roadmap/PLAN.md:95-96). Those conflicting operational instructions can cause the remaining Desktop change or a later promotion review to follow the old cadence and threshold rather than the ratified policy; update the live summaries alongside this status amendment.

Useful? React with 👍 / 👎.

[kyle-sexton]

#	Finding	Classification	Evidence	Reacted
1	Live summaries still encode the pre-rebind policy: Brief defines hourly drain + ≥14d predicate, operator-decision table directs an hourly preset, parent roadmap scopes Phase I as hourly	VALID — surfaced to operator	Confirmed at head a0c407e7: Brief (PLAN.md ~L12) "hourly C2 drain" and predicate "≥14 days"; operator-decision table Schedule row (~L29) "Hourly preset"; parent ladder-climb-roadmap/PLAN.md:95–96 still hourly. The Phase 4 amendment ratifies 15-min/7-day, so the live summaries conflict	👍

Disposition (safe tier): not editing and not resolving. This PR carries needs-human; rewriting the authoritative Brief + operator-decision table, and editing a second document (ladder-climb-roadmap/PLAN.md) not in this PR's diff, is beyond the stated docs-status-amendment scope — an operator/design call, not a clear branch-owned fix.

Note: this and the L273 thread are two loci of one operator decision — propagate the ratified 15-min/7-day rebind into the live Brief / operator-decision table / parent roadmap / Phase 2 detector, or deliberately keep those as design-time records with the Phase 4 amendment layered on top. Surfaced to @kyle-sexton to decide and, if propagating, amend; resolve this thread to ratify keeping it as-is.

[claude]

Claude encountered an error after 1s —— View job

---

I'll analyze this and get back to you.