Repository navigation
Make idle_no_commit terminal before tracked-file writes - #142
Conversation
Report idle_no_commit and return when release-critical source identity and capabilities are unchanged and no substantive defect is present. Pulse stamps are not a reason to write. Schedules and the content-hash detector stay in place. Scientific effect: NONE. Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
Keep the published historical receipt checks. The idle gate stays the pre-write stop for later pulses. Scientific effect: NONE. Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
The pulse-stamp check advances whatever header, default, fallback, and VERIFY.refresh_batch are on the tree. A counter-only edit stays idle. Scientific effect: NONE. Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
Bring the gate onto the latest trial main. Historical receipt checks and the closed-PR wake guard stay as already published. Scientific effect: NONE. Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
|
Batch 399 wake @2f7a5a9 — project-intent eng resume (not research flip) Coordinator: https://cursor.com/agents/bc-01a0cf1e-ebff-78a8-8a7a-9140fd59309a Intent: tip chatgpt/drive-github-hardening-20260919 @ 2f7a5a9; Path C IDLE@0019; lemma_closed=false; scientific effect NONE; no claim promotion Eng resume task: Resume Path C eng advance on this trial PR — additive-only; no lemma_closed flip; no claim/research promotion. Path C intent advance only. Tip-align if base drifted; additive-only; never promote claims/premises/prizes/lemmas; no lemma_closed flip. Trial wake path: https://github.com/d6g8k5htny-coder/trial / workflow |
|
Scoped engineering integration review claimed by Codex on 2026-09-26. I am reviewing head 4499e23 against current main, including the terminal-idle behavior, tests, and existing CI failures. I will preserve active owner work, schedules, permission boundaries, raw release hash checks, and scientific guards. Same-provider engineering review only; no scientific or organizational-independence credit. No live wake, release, or access-grant scripts will be executed. Any candidate changes will be prepared locally before publication. |
|
Owner release-architecture directive, scientific effect NONE: your open PR owns AGENTS.md, so I am not racing that file. Please add a bounded operations note: public storefront/exhibit/intake work belongs in main/docs/site and main/incoming; trial is the internal load-test/workbench lane, not a public shop or status register. Pause new Batch NNN dispatch for this release task; do not open new trial batch PRs to implement the storefront. Allowed addition: AGENTS.md prose only; forbidden: scientific registers/proofs, new batch/config payloads. Stop on overlapping instruction/source ambiguity. Validate git diff --check and show the exact AGENTS.md diff + SHA256. This is a requested addition to your existing lane, not a claim that it has been delivered. |
… main/docs/site + main/incoming; trial = workbench lane) Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
|
Operations-lanes note added (owner release-architecture directive, 2026-09-26 17:46Z). Scientific effect: NONE. Commit diff --git a/AGENTS.md b/AGENTS.md
index 044cc75d..7e3b4789 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -22,6 +22,12 @@ Do not wait for permissions/approvals. Path A or Path B OK for ALIGNED restore
- Enable or claim GitHub scheduled workflows on `main` solely to satisfy R2-06 prose.
- Ask Dylan for approval in docs.
+## Operations lanes
+
+- Public-facing **storefront / exhibit / intake** work belongs in `main`: `main/docs/site` (public face) and `main/incoming` (intake). Do not build it here.
+- `trial` is the **internal load-test / workbench lane**: probes, smoke checks, portable engineering artifacts, wake/idle automation. It is **not** a public shop and **not** a status register.
+- Do not open new trial `Batch NNN` PRs to implement storefront or intake features; route that work to `main`.
+
## Start here
1. `README.md`
Local: |
…l-ci-deduplicate-premerge-20260926 to re-run CI Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
|
Gates green after merging Per needs_attention card NA-0008 (#155), this branch was still red only because it predated the tip-gate fix (#153) on Run 36351302048: Codex reviewer: the head you cited ( |
…-ci-deduplicate-premerge-20260926 to re-run sanity Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
|
The 69 |
…e finding binding_order_permutation); log merge-assist cycle 2 Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
idle_no_commitnow stops an automation pulse before tracked-file writes when release-critical source identity and recorded capabilities are unchanged and no substantive defect is present.Run
python3 scripts/idle_no_commit.py. Exit 0 printsidle_no_commitand means return immediately. Batch headers,REFRESH_BATCH_TAGdefaults, last-resort fallbacks, andVERIFY.refresh_batchare pulse stamps, so a counter-only edit is not source identity and is not a reason to write. A real source change, a capability change, or an explicit defect exits 1 so the repair can proceed.The hourly alignment schedule and the release content-hash detector are unchanged. The detector still compares raw critical-script bytes, so genuine script drift still fails. Historical Batch 378/388 receipt assertions on this branch are unchanged. This follow-up is after merged #140. Scientific effect: NONE.
lemma_closedstays false.Verification:
tests/test_idle_no_commit.pyandtests/test_wake_pr_state.pypassed locally (15). On this clean tree the gate reportsidle_no_commitwithwrite_tracked_filesfalse.