Repository navigation
fix: mint the repo and clear the five red estate checks (stacked on #4) - #6
Merged
Merged
Conversation
…finition The harness accepted any diagnostic in a reject fixture that mentioned the file and said "when checking that". An earlier, unrelated type error therefore passed as the expected rejection before Agda ever reached the claim the fixture exists to refute, and a fixture with no `invalid` definition at all passed too. The first location Agda reports in the fixture must now lie inside the `invalid` definition: from its first clause (not the type signature) to the last clause or continuation line before the next top-level declaration. A fixture with no `invalid` definition fails loudly. Validated on Agda 2.6.3: all five fixtures still pass, each at its `invalid` line. Negative controls run against the old and new script: - earlier unrelated error: old PASS, new FAIL (line 7, outside 10-10) - no `invalid` definition: old PASS, new FAIL - error on a continuation line of `invalid`: old PASS, new PASS Reported by CodeRabbit on resource-algebra#4. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
…ectly Found while minting this repository, which was created with GitHub's "use this template" (its first commit carries rsr-template-repo df09a6c's tree byte for byte) and never minted. Every fix below was exercised by the mint in the next commit. 1. RSR_NON_INTERACTIVE never worked. It shadowed `read` with a no-op for the whole recipe, so (a) any prompt the environment did not answer left its variable unset and `set -u` killed the mint at OPENSSF_BP_ID, after docs/ had already been overwritten, and (b) every `while read -r` loop inherited the no-op and spun forever on an unbound `file`. Prompts now go through `ask`, which keeps the environment's value (or empty) and leaves the builtin alone. The e2e test drives only the interactive path, which is why nothing noticed. 2. The token pass rewrote just's own interpolation. In Justfile and *.just every double-brace expression is a just variable; the lowercase project token is spelled like just's interpolation of `project`, so seven Justfile lines, the contractile Justfile and container.just's image tag froze into `resource_algebra`. just files are now skipped by the token pass; the self-name pass still rewrites them. 3. The mint left CLAUDE.md and the repository map describing the template. The arrival pack named this repo the estate's `foundation` with the template's uuid and clade; the map kept archetypes/, which the mint deletes, so Estate Rules' map-freshness step failed in every minted child. Both are regenerated after the deed is final. 4. The self-name pass rewrote a detector. The deed's state-validation `:reject-if-contains ("rsr-template-repo")` became the child's own slug, so it rejected every legitimate state file of the child. Lines carrying `reject-if-contains` are now protected like provenance lines. The same change belongs upstream in rsr-template-repo. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
This repository was created with GitHub's "use this template" from rsr-template-repo df09a6c (its first commit carries that commit's tree byte for byte) and was never minted. It therefore still declared itself the template: rsr-template-repo_chora.deed with the template's uuid, clade and `foundation` pipeline position, 87 files of unfilled placeholder tokens, a CODEOWNERS assigning no owner, and a MAINTAINERS naming nobody. Four estate checks were red on main for exactly that: OpenSSF Compliance, Central Estate CI/CD Audit, Dogfood Gate (Canon lockstep) and, behind the root-shape failure, Estate Rules. This commit is the unedited output of the fixed recipe in the previous commit, run non-interactively: RSR_NON_INTERACTIVE=1 PROJECT_NAME="Resource Algebra" \ REPO=resource-algebra OWNER=hyperpolymath \ AUTHOR="Jonathan D.A. Jewell" AUTHOR_EMAIL=j.d.a.jewell@open.ac.uk \ AUTHOR_ORG="The Open University" \ PROJECT_DESCRIPTION="The parametric grade algebra shared by the hyperpolymath type family" \ just repo-init Values are the ones CITATION.cff and the SPDX headers already state. The mint reported: all placeholders replaced; deed renamed to resource-algebra_chora.deed; uuid 241f1c29-0838-5033-b654-7f4c5e279ebc (uuid5 of github.com/hyperpolymath/resource-algebra); CLAUDE.md and the repository map regenerated; `just verify` PASS. What the recipe cannot know (true template provenance, birth date, clade) follows in the next commit. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
…d is given Exposed by the previous commit's mint. With OPENSSF_BP_ID blank, the recipe deleted every line of docs/governance/TEMPLATE-STANDARDS-AUDIT.adoc holding the OpenSSF token. That line is one cell of a six-column table row, so the row lost a cell, asciidoctor reported "dropping cells from incomplete row detected end of table", and Estate Rules' render gate failed in every child minted without an OpenSSF id. The document names the README's badge tokens rather than using them, so the row now spells them as plain names (OPENSSF_PROJECT_ID, OWNER, REPO). Neither the substitution pass nor the placeholder gate touches them, and the recipe no longer edits that file. Verified: the file renders clean under asciidoctor 2.0.26 --failure-level=WARN, and scripts/check-adoc-renders.sh passes on the whole tree. The same change belongs upstream in rsr-template-repo. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
What the mint cannot know, corrected by hand in the deed and projected into CLAUDE.md by `just claude-md`: - Provenance. The mint ran on this repository's own history, so it recorded a resource-algebra commit as the template parent. The real parent is rsr-template-repo df09a6c on main: this repository's first commit carries that commit's tree (2c01b6f) byte for byte. (variant :parent-pin) and (provenance :template-commit/-tree/-branch) now say so. - Birth. The repository was created 2026-10-12, not at the 2026-10-14 mint: (lineage :born), (status :since) and its history entry. - Clade fv, Formal Verification & Proofs, chosen by the owner. The template leaves the clade UNASSIGNED because choosing is a human act. - (ecosystem) :position-type, :purpose and :not, transcribed from the README's own Overview and "What this is not" rather than invented. CLAUDE.md's generated region now states this repository's identity (uuid 241f1c29-..., clade fv, born 2026-10-12) instead of the template's; .machine_readable/arrival-pack/verify.sh reports it in sync with the deed. Registration in gv-clade-index remains the owner's act and is not done. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
resource-algebra#4 appended the owner's rule (put every question that needs the owner's decision through the AskUserQuestion selector) to CLAUDE.md as a hand-authored section, and CodeRabbit flagged it as out of scope for #3. The owner ruled (2026-10-14) that it belongs in the estate arrival pack: it governs how every agent asks the owner, in every repo. The rule now lives in the estate-common block of .machine_readable/arrival-pack/arrival-pack.ncl, wording unchanged, and CLAUDE.md carries it inside the generated region instead of below it. .machine_readable/arrival-pack/verify.sh reports the pack in sync. The canonical copy of arrival-pack.ncl is rsr-template-repo's; the same edit goes upstream so every repo receives the rule on regeneration. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
Estate Rules failed at its second step on main (root shape), so none of
its later steps had ever run here. With the root fixed, the render gate
was the next failure. All 18 steps now pass locally, run as CI runs them.
- PROOF-STATUS.adoc moves to docs/status/PROOF-STATUS.adoc, the
template's canonical location; the allowlist records PROOF-* at the
root as drift the April 2026 root cleanup relocated. It replaces the
unused template stub that sat there. README links follow.
- resource-algebra.agda-lib stays at the root, as a justified REQUIRED
allowlist entry: Agda finds the library file by walking up from each
source file, consumers register its path, and Agda convention keeps it
at the project root. Moving it would break consumers' registrations.
- README.adoc had two level-0 titles (lines 4 and 8), which asciidoctor
rejects outside a book; the duplicate title block is removed. Below the
owner's "RSR template sections below; swept by `just repo-init`"
marker, the sections the template itself marks for deletion after a
mint are removed (its "this file is a template" admonition, "What this
is", "Quick start"), and the AI-install lines the self-name pass had
made false ("Create a new repository from .../resource-algebra") take
the canonical child wording from standards'
3-practice/provisioning/templates/README-ai-install.adoc.tmpl. The
[[ai-install]] section that `just validate-ai-install` requires stays.
- docs/architecture/REPOSITORY-MAP.adoc regenerated (`just repo-map`).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
The Central Estate CI/CD Audit stopped at its first gate (Required Files) on main, so its other 25 gates had never run here. Run locally from cicd-suite 0da816c, the pin main-estate-audit.yml uses, 25 of 26 passed after the mint; the Academic Referencing Gate failed with "CITATION.cff is missing the required top-level 'message' field". `message` is a required key of the CFF 1.2.0 schema. The wording is the template's own. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
Governance / Actions lockfile verify was red on main. At the pinned standards commit da2c748, governance-reusable.yml installs github/gh-actions-lock UNPINNED, so a new release (cli v0.2.0) changed the verdict for every caller. Its staged verifier accepted the seven reusable-workflow "stale" findings as coverage, as designed, but then failed on two `sha-as-ref` warnings for codeql.yml's bare-SHA github/codeql-action pin. Standards main already carries both cures (standards#1120 pins the tool to v0.1.6, and update-actions-lock.sh treats sha-as-ref as advisory). A pinned caller never receives a cure until its pin moves, so this moves it. - The six standards reusable calls now pin ae36881 (standards main, 2026-10-14), rewritten by standards' own scripts/propagate-workflow-pins.sh --fix, whose ancestry check passed. - actions.lock follows by hand, as check-lock-sync.sh prescribes for job-level reusable refs: the six workflow entries are re-keyed, the hyperpolymath/standards dependency record is rebuilt for ae36881 with the actions its reusables use at that commit, and the three records it newly needs (actions/upload-artifact@cf430e0 v7.0.2, dtolnay/rust-toolchain@e2a55d2, Swatinem/rust-cache@6323deb) are copied verbatim from standards' own actions.lock. Four records only the old commit used are pruned; record order matches the generator's. Verified: zero dangling edges (GitHub's startup rule), and both scripts/check-lock-sync.sh and standards main's copy pass. The gh-actions-lock --verify-local half runs only in CI. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
Contributor
|
Important Review skippedToo many files! This PR contains 161 files, which is 61 over the limit of 100. To get a review, reduce the PR to 100 files or fewer by splitting it into smaller PRs or changing its base branch. Upgrade to a paid plan to raise the limit. This review couldn't start because sufficient usage credits or metered capacity aren't available. Add credits or update usage-based reviews in the billing tab, then retry. ⚙️ Run configuration
⛔ Files ignored due to path filters: 2
📒 Files selected for processing: 161
You can disable this status message by setting the
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
In non-interactive mode every other prompt falls back to the variable of its own name (OWNER, AUTHOR, FORGE, ...), but the service name, port and registry read only `_SERVICE_NAME`, `_PORT` and `_REGISTRY`, so SERVICE_NAME=... in the environment was silently ignored. They now fall back to SERVICE_NAME, PORT and REGISTRY like the rest, and the prompts show the effective default, because PORT is a common shell variable and a blank answer must not pick it up unannounced. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
6 tasks done
hyperpolymath
added a commit
to hyperpolymath/rsr-template-repo
that referenced
this pull request
Oct 14, 2026
…y; standards re-pin (#237) ## Summary Root-cause fixes for the estate checks that were red on [resource-algebra#4](hyperpolymath/resource-algebra#4). That repo was created with GitHub's "use this template" from `df09a6c` and never minted. Minting it ([resource-algebra#6](hyperpolymath/resource-algebra#6)) exposed defects in this template's mint recipe, and showed that every child inherits a broken standards pin from here. These are fixed at source so new repos don't inherit them. ## Changes - **`fix(repo-init)`**: six recipe defects, each exercised by real mints. - `RSR_NON_INTERACTIVE` stubbed `read` for the whole recipe. Unanswered prompts then died under `set -u`, and the substitution loops never ended. Prompts now go through `ask`, which reads the environment. Service name, port and registry now honour their environment variables too. - The token pass rewrote `just`'s own interpolation of `project` in Justfile and `*.just`. `just` files are now skipped by that pass. - CLAUDE.md and the repository map were left describing the template. Both are now regenerated from the child's deed and tree. - The self-name pass rewrote the deed's `reject-if-contains` detector into the child's own name. - A blank OpenSSF id deleted a table cell in `TEMPLATE-STANDARDS-AUDIT.adoc`, breaking the render gate. - `identity.deed.in` gains `:pipeline-position` and `:chain` as `UNASSIGNED`. - **`test(repo-init)`**: `tests/workflows/non_interactive_mint_test.sh`, wired into Estate Rules with a checksum-verified `just` 1.43.0. No workflow ran `repo-init` before; `tests/e2e` runs nowhere in CI. - **`docs`**: four documents told agents never to use `RSR_NON_INTERACTIVE`. They now describe the working environment-variable path. The guide's prompt table gains the missing ORCID row and a variable column. - **`docs(arrival-pack)`**: the owner's "Asking the owner" rule joins the estate-common block, per the owner's ruling of 2026-10-14. - **`fix(ci)`**: the six standards reusables re-pinned from `da2c748` to standards main `ae36881`, with `actions.lock` re-keyed and closed. At `da2c748` the governance reusable installs `gh-actions-lock` unpinned, and v0.2.0 fails every caller on `sha-as-ref`. Standards main pins v0.1.6 and treats `sha-as-ref` as advisory. - **`fix(archetypes)`**: the julia-library lock fragment moves to format v0.0.3. [#236](#236) moved the spine lock to v0.0.3, so `archetype_lock_merge_test.sh` fails on `main` and Julia mints stop at the merge. - **`docs(changelog)`**. ## RSR Quality Checklist ### Required - [x] Tests pass: every Estate Rules step and the key gates run locally on this tree (below) - [x] No banned language patterns - [x] SPDX license headers present on all new files - [x] No secrets, credentials, or `.env` files included ### As Applicable - [x] Documentation updated - [x] `CHANGELOG` updated under [Unreleased] / Fixed ## Testing - **Template mode.** All Estate Rules steps pass on this tree, including the new test. `archetype_lock_merge_test.sh` fails on `main` without the fragment fix; it was reproduced on `e2c87e6`. Placeholders, root shape, render, lock sync (repo and standards main) and the arrival-pack check all pass. - **The new test, watched failing.** On the old recipe it fails twice over: the `read()` shim, and "PROJECT_NAME: unbound variable". It passes on the new recipe with `just` 1.43.0. - **A full non-interactive mint of this tree** with only the five required answers (`mint-check`) exits 0. The minted child passes placeholders, root shape, render (158 files), arrival pack and map freshness. It has no underscored names in `just` files and keeps its detector. - **resource-algebra** was minted with the same recipe. There, all 18 Estate Rules steps and all 26 Central Estate Audit gates (cicd-suite `0da816c`) pass locally. Not verifiable locally: the governance reusable's `gh actions-lock --verify-local` half. CI is the evidence for it. ## Not fixed here - `scripts/gen-repo-map.sh` breaks on an apostrophe in a `root-allow.txt` comment (`xargs: unmatched single quote`). - `docs/AI_INSTALLATION_GUIDE.adoc` and the README tail are written for the template, and a minted child keeps them verbatim. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi --- _Generated by [Claude Code](https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi)_ --------- Co-authored-by: Claude <noreply@anthropic.com>
hyperpolymath
marked this pull request as ready for review
October 14, 2026 11:12
#4 merged with five more commits that overlap this branch, and #5 bumped actions. Resolved in favour of the owner-merged content, keeping what main does not yet carry (the mint itself: 87 placeholder files, the template-only paths, provenance): - check-agda.sh: both hardenings. main's `-- EXPECT:` message check and unit-test run, and this branch's rule that the first error must land inside `invalid`. All 8 fixtures, including main's three new ones, pass both on Agda 2.6.3; the unit tests pass. - Deed: clade fv (the owner's answer) and this branch's provenance, variant and rsr-profile; main's phase `active`, ecosystem (library, kernel, chain), anchor kind/domain and unique strength. `:coordination standards` restored (main's ecosystem dropped it). - PROOF-STATUS, CITATION message, CODEOWNERS comment, allowlist lines: main's. The duplicate .agda-lib allowlist entry is dropped. - ROADMAP: the minted stub. main's file was still the template's own journal ("A repo minted from the template does not inherit it"). - actions.lock: main's upload-artifact v7.0.2 plus the cf430e0 record the standards@ae36881 record needs. - CLAUDE.md and the repository map regenerated from the merged tree. main removed #4's hand-appended "Asking the owner" section (6c61289); the rule stays only in the generated estate-common block, per the owner's ruling that it is estate doctrine. Verified on the merged tree: Estate Rules 18/18 steps, all 26 Central Estate Audit gates (cicd-suite 0da816c), the placeholder gate, both lock-sync gates, the arrival-pack check, and scripts/check-agda.sh. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
|
6 tasks done
hyperpolymath
added a commit
that referenced
this pull request
Oct 14, 2026
## Summary [resource-algebra#6](#6) was merged at `b0e4189`, before its last two commits were pushed. Main (`768068f`) therefore still fails two governance jobs. This PR carries those two commits, rebased onto main. | Check on main | Cause | Fix | |---|---|---| | governance / Actions lockfile verify | standards `ae36881` installs `gh-actions-lock` v0.1.6, which refuses lock format v0.0.3 ("newer than this binary supports") | re-pin the six standards reusables to `934b879`, whose governance job pins v0.2.0 | | governance / UUID v7 conformance | the arrival pack printed the deed's uuid5 as a hex literal into CLAUDE.md; `check-uuid-v8.sh --strict` rejects non-v8 hex | print the deed's typed literal `#u5"github.com/hyperpolymath/resource-algebra"` instead | ## Changes - **`fix(arrival-pack)`**: `extract.sh` takes `:repo-uuid` as the typed `#u<n>"…"` literal, as the deed writes it, rather than as hex. A deed whose uuid is not a typed literal stops generation (exit 2) instead of printing an empty field. CLAUDE.md is regenerated. - **`fix(ci)`**: the six `hyperpolymath/standards` reusables are re-pinned to `934b879` (standards main). `actions.lock` is re-keyed with standards' `propagate-workflow-pins.sh`, using the relock step proposed in [hyperpolymath/standards#1254](hyperpolymath/standards#1254). The lock gains the two records the new pin reaches (`denoland/setup-deno`, `oven-sh/setup-bun`). ## RSR Quality Checklist ### Required - [x] Tests pass: `bash scripts/check-agda.sh` (unchanged by this PR) - [x] No banned language patterns - [x] No banned functions - [x] SPDX license headers present on all new/modified source files - [x] No secrets, credentials, or `.env` files included ### As Applicable - [x] Documentation updated (CLAUDE.md regenerated, not hand-edited) ## Testing On this tree: - `standards/scripts/check-uuid-v8.sh --strict .`: exit 0. On main it fails on `./CLAUDE.md`. - `scripts/check-lock-sync.sh`: exit 0, with 0 dangling edges and a lock key for every workflow. Standards main's copy also passes. - `.machine_readable/arrival-pack/verify.sh`: CLAUDE.md matches the generator. I can't run the `gh actions-lock --verify-local` half of the lockfile job locally, so CI is the evidence for it. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi --- _Generated by [Claude Code](https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi)_
hyperpolymath
added a commit
to hyperpolymath/standards
that referenced
this pull request
Oct 14, 2026
…n bumpers (#1254) ## Why `scripts/apply-workflow-pins-remote.sh`, the scheduled pin applier, fetches and commits only `*.yml`. The local `propagate-workflow-pins.sh` has the same gap. Both re-point `uses: hyperpolymath/standards/...@SHA` and leave `actions.lock` keyed to the old commit. `check-lock-sync.sh` rejects that state, and its own header records GitHub refusing such runs at startup. The ledger census in `apply-workflow-pins.yml` counts 200 of 368 governance callers with an `actions.lock`. A fix-mode run would therefore have opened PRs that stop those repos' reusable-calling workflows. This came up while fixing a stale `da2c748` pin, whose unpinned `gh-actions-lock` fails Actions lockfile verify. That fix landed in [resource-algebra#6](hyperpolymath/resource-algebra#6) and [rsr-template-repo#237](hyperpolymath/rsr-template-repo#237), and both needed this relock done by hand. ## What - **`scripts/relock-standards-pin.sh`** (new) moves a consumer's lock to the target standards commit: - re-keys the per-workflow entries; - rebuilds the standards dependency record from what the reusables called with a full pin use at the target; - copies the records those need verbatim from standards' own lock; - prunes only records the old standards record alone kept alive; - keeps `gh-actions-lock`'s record order. - **Short pins.** A short pin of 7 or more hex that prefixes the target is current. This matches the applier, which counts it FRESH and leaves it as written. It keeps its own short-keyed record, untouched. - **It is fail-closed.** It leaves the lock untouched and exits 1 when: - a workflow pins another standards commit, or a pin under 7 hex; - a reusable is absent at the target; - the old record is missing; - a needed record is in neither lock; - a workflow's standards entries would not name exactly the refs it pins. - **The applier** fetches `actions.lock` with the workflows and relocks it in the same verified commit. A repo whose lock can't follow is reported `LOCK-REFUSED` and not landed. In fix mode it fetches the target commit if the CI checkout lacks it, and a new known-answer control fails the run if the relock is missing. - **The propagator** relocks before staging and reports `RELOCKED` / `LOCK-REFUSED`. On a refusal it restores the workflows it rewrote, uncommitted edits included, stages nothing, and leaves the branch alone. Any refusal makes it exit 1. - **Ported:** the frozen job-name fix that #1255 and #1256 also carry. Self Test is red on main without it. ## Evidence - **Known answers.** On resource-algebra and rsr-template-repo, re-pinned from `da2c748` to `ae36881`, the relock reproduces byte for byte the locks verified by hand there. Both pass `check-lock-sync.sh`. resource-algebra's lock relocked to `934b879` passes the governance gate's own `update-actions-lock.sh --verify-local` with gh-actions-lock v0.2.0. - **`scripts/tests/relock-standards-pin-test.sh`**, offline, 16 checks: - the known answer, and idempotence; - eight refusals that leave the lock byte-identical; - the no-lock no-op; - short pins at 7, 12 and 39 hex, each passing `check-lock-sync.sh`. - **`scripts/tests/propagate-workflow-pins-test.sh`**: 29 of 29. - **Mutants**, all killed: - from the first round: no iteration, no pruning, no record copying, no stale-pin refusal, over-pruning, no sorting; - from the review round: exact-match only, every pin feeding the record, no 7-hex minimum, no consistency guard, no restore, no early return. - **Existing suites.** `check-lock-sync-test.sh` 13 of 13, and the applier's `--self-test`. The review round built on CodeRabbit's autofix commits (`cccec83`, `5fd1373`) and kept their tests. Not exercised: a live fix-mode applier run. It needs the dedicated App credential, and it is the owner's dispatch. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi --------- Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Stacked on #4. This branch starts at #4's head, so until #4 merges its diff includes #4's commit. Merge #4 first and this PR shrinks to its own eleven commits.
The five checks red on #4 were all red on
maintoo. Four had one root cause: the repo was created with GitHub's "use this template" from rsr-template-repodf09a6c(its first commit carries that commit's tree byte for byte) and was never minted. It still declared itself the template. Minting it exposed six defects in the template's own mint recipe, fixed here and to be sent upstream.{{…}}tokensmessageresource-algebra_chora.deedPROOF-STATUS.adoc,resource-algebra.agda-lib; README with two titlesda2c748installsgh-actions-lockunpinned (v0.2.0)It also fixes the CodeRabbit nit on #4 (
check-agda.sh), and moves #4's hand-written "Asking the owner" CLAUDE.md section into the generated arrival pack, as the owner ruled.Changes
fix(check-agda): a rejection must land inside the fixture'sinvaliddefinition. Negative controls: an earlier unrelated error and a fixture with noinvalidboth passed the old script and fail the new one.fix(repo-init)×2, template defects found by minting:RSR_NON_INTERACTIVEshadowedread, so the automated mint crashed at an unset variable or looped forever.just's own{{project}}interpolation.reject-if-containsdetector.chore: mint: uneditedjust repo-initoutput.fix(deed): true template provenance, birth date 2026-10-12, cladefv(owner's choice), and IS / IS-NOT taken from the README.docs(arrival-pack): "Asking the owner" now sits in the estate-common block.fix(estate-rules): PROOF-STATUS moves todocs/status/; the.agda-libgets a justified allowlist entry; README template leftovers go; the map is regenerated.fix(citation): the CFF 1.2.0messagefield.fix(ci): the six standards reusables re-pinned toae36881with standards' ownpropagate-workflow-pins.sh;actions.lockre-keyed and closed.docs(changelog).RSR Quality Checklist
Required
bash scripts/check-agda.sh, core proofs plus all 5 rejections.envfiles includedAs Applicable
(status)/(maturity)updated(ecosystem)updatedCHANGELOGupdatedTesting
Every step of the five failing workflows was run locally on this tree with the CI-pinned tools (Agda 2.6.3 locally, Nickel 1.17.0, asciidoctor 2.0.26):
cicd-suite@0da816c. Onmainit stopped at gate 1.actions.lock: zero dangling edges.Not verifiable locally: the governance reusable's
gh actions-lock --verify-localhalf, the deed and k9 ecosystem validators, and the RSR oracle. CI is the evidence for those.Open for the owner: register
resource-algebrain gv-clade-index (the owner's act).docs/AI_INSTALLATION_GUIDE.adocis still the template's text.🤖 Generated with Claude Code
https://claude.ai/code/session_01LiQzWJdcPfRU6YMDGRG8Xi
Generated by Claude Code