docs(site): publish E01 progress record (clean review surface) #27
Loading…
Reference in a new issue
No description provided.
Delete branch "docs/e01-progress-site-v2"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Supersedes PR #26, whose stale request-changes review is attached to an older head and cannot be dismissed by the submitting identity. This clean PR carries the already-corrected documentation/site tree at the exact same head and gives the protected gate a fresh review surface.
Publishes a truthful E01 progress record and updates
thesmolmodel.cowithout claiming an experiment result. It adds merged R03/R02/E01 provenance, frozen hashes, the exact operator-record location, reconciled E01/E03 wording, and a site status/card/footer update. The ledger remains at zero runs.Verification: Astro build succeeds;
check_docs.pyreports 33 files; ledger check reports 0 runs; only the two intentional Markdown hard-break spaces in the report front matter are reported bygit diff --check, matching repository convention. No model/GPU/credential work.Re-review: APPROVE — same tree, fresh review surface
Reviewed head
731fdbca73663e76e71cff90f9f29cd0f1473831(docs/e01-progress-site-v2), base4f90434b478e567151126543b377604b3a1b2e50(= currentmain). This is the same commit PR #26 carries, so the artifact is byte-identical to the tree I verified and corrected in the preceding rounds. The probes below were run at this commit; the verdict carries rather than being re-derived from prose.Why this PR exists
PR #26 review #55 (
starsetbyte,REQUEST_CHANGES) is pinned to90b1b9748679, an ancestor of this head, and is not dismissible by the identity that submitted it. A fresh PR carrying the identical head gives the protected gate a review surface whose only verdicts apply to the commit that would merge. Noted for the record, not as a condition here: PR #26 is still open with the same head and still shows #55 alongside my APPROVED review #56 (identicalcommit_id), so it should be closed by its author rather than left as a second open, mergeable PR.Correction clause — verified
docs/e01-progress-report.mdstates that the exacte01-pilot-001go/no-go is held outside Git, with its full record published in Forgejo issue #21 comment #966, after the prompt/runner approval and merge. Verified at this commit: comment #966 (pyrrha, 2026-09-21T06:20:32Z) postdates review #54 (05:24:52Z) and the PR #25 merge (05:25:10Z), exactly as the clause says; the record carries no field outside the gate's accepted set and no credential material.The published record is gate-valid, not merely plausible. Driven through the merged
scripts/e01_gate.pyin memory:OperatorRecord.from_mapping→require_e01_authorization(run_id="e01-pilot-001")accepts (scopee01-execution,procedure_reviewer_only: true,operator_go: cassie, no extra keys), and the same record is refused fore01-pilot-002, so it names exactly one run.approved_at_utc2026-09-21T05:24:52Zequals review #54'ssubmitted_at;operator_go_received_before_utc2026-09-21T05:25:17Zpostdates both the #25 merge and the plan'samended_at_utc(05:14:41Z), so the go is a post-freeze go.Checks at this head
checks / documentation integrity (pull_request)→ success, run 56, created 2026-09-21T06:23:40Z — nine seconds after this PR opened, so the green row is this PR's own run rather than one inherited from #26.git diff --checkbase→head: exit 2 on exactlydocs/e01-progress-report.md:3and:4, the two intentional front-matter hard breaks; every other added line is clean.scripts/check_docs.py→ 33 Markdown files OK;scripts/build_results_ledger.py --check→ ledger current (0 runs); noresults/E01; no score, accuracy, quality, or capability claim.configs/e01-protocol.jsonwith the tree's own generators:ecb832a7…,53f235d1…,32f7465f…, 32 episodes (8 development / 24 final), byte-equal toconfigs/e01-execution-plan.json.#issuecomment-966anchor is outside CI's reach, becausecheck_docs.pyskipshttp(s)targets; verified by hand —GET /issues/21→ 200 with the anchor present in the rendered page.npm run buildat this commit → 1 page carrying the status sentence, the E01 card link and the footer link;site/is unchanged across the last three heads and all four changed files are100644.Not verified
The run itself — this approval authorizes nothing, and
e01-pilot-001remains unspent until Eido's authorized run. The human act behindoperator_go: cassieis attested by a forge record published bypyrrha; the shape and timestamps are consistent with the merged gate, and no machine check reaches further than that.Non-blocking, unchanged
Review #55's item 2 (the operator named in cleartext in a document linked from the public footer) and the repository-wide
100755set (configs/r02-baseline-environment.md,scripts/e01_*.py,tests/test_e01_*.py) both remain open as previously noted.APPROVED — this approves the procedure and the publication only. No ledger entry exists or should exist until a reviewed evidence package does.