issue-89: Validate install:local from a scratch clone (fresh machine) #97

Merged

david merged 2 commits from feature/issue-89/validate-install-local-from-a-scratch-clone-fresh into main

2026-08-28 01:25:12 +00:00

david commented

2026-08-28 01:25:10 +00:00

Owner

89: Validate install:local from a scratch clone (fresh machine)

Implements 89.

This pull request was generated automatically by pi-loop and opened as a
Draft for human review. Do not merge without reviewing the changes.

Generated in 27m 33s.

Review findings

Now let me verify the commit-message convention claim against the actual git history (read-only):

Confirmed: The commit is on the unpushed branch feature/issue-89/... (1 commit ahead), and recent history consistently uses the bare <N>: format for commits, while PR titles use issue-<N>: (e.g., PR #96). Now let's apply the fix:

Medium finding — The safest single-pass fix is to rephrase (filing an issue would modify remote state that isn't captured in the working tree diff, and it's not appropriate for this stage):

The only gate in package.json is test. Re-running it to confirm nothing regressed:

All 35 tests pass. Final verification of the working-tree change:

All gates green and both decisions are made:

Medium — resolved by rewording Observation 3 (filing a remote issue would mutate state outside the captured working-tree diff; the reword was the finding's explicitly offered alternative).
Low — left unresolved: amending 0febc4f requires rewriting git history, which this stage is prohibited from doing (no commits/stages/pushes), and reconciling AGENTS.md with the bare <N>: convention is a repo-owner policy call (practice is split — commits use <N>:, PR titles use issue-<N>: per PR #96).

Review Findings — 89

Summary

Docs-only change adding a "Results — issue #89" validation record to docs/IMPLEMENTATION_PLAN.md. Remediation: reworded Observation 3's false claim ("Flagged for a follow-up issue" — no such issue exists) to "Should be flagged for a follow-up issue (none filed as of 2026-08-28)". The Low commit-message-format finding remains unresolved: amending 0febc4f requires rewriting git history (prohibited in this stage), and reconciling AGENTS.md with the bare <N>: convention is a repo-owner policy decision. Gates re-run after the fix: npm test → 35/35 pass; build N/A and lint N/A (no such scripts in package.json; docs-only change).

Critical

(none)

High

(none)

Medium

docs/IMPLEMENTATION_PLAN.md:117 — Observation 3 asserts the README-vs-manifest victorialogs mismatch was "Flagged for a follow-up issue", but no such issue exists in the Forgejo project (searched victorialogs → 0 issues; manifest/README → only unrelated issues). The claim is currently false and will mislead readers into thinking tracking exists. Suggested fix: file the follow-up issue and reference its number here (e.g. "flagged as #"), or reword to "should be flagged for a follow-up issue".

Low

docs/IMPLEMENTATION_PLAN.md:96 — Commit 0febc4f message is 89: Validate install:local from a scratch clone (fresh machine), but AGENTS.md requires the issue-<N>: <summary> format for issue-driven work (e.g. issue-42: add user auth). It matches this repo's recent history (commits 70–88 all use bare <N>:), so it is consistent with practice but deviates from the letter of the standard. Suggested fix: amend to issue-89: Validate install:local from a scratch clone (fresh machine) before merge, or reconcile AGENTS.md with the established <N>: convention.

## 89: Validate install:local from a scratch clone (fresh machine) Implements [89](https://git.excelera.net/david/pi-extensions-and-skills/issues/89). > This pull request was generated automatically by pi-loop and opened as a > **Draft** for human review. Do not merge without reviewing the changes. Generated in 27m 33s. ## Review findings Now let me verify the commit-message convention claim against the actual git history (read-only): Confirmed: The commit is on the unpushed branch `feature/issue-89/...` (1 commit ahead), and recent history consistently uses the bare `<N>:` format for commits, while PR titles use `issue-<N>:` (e.g., PR #96). Now let's apply the fix: **Medium finding** — The safest single-pass fix is to rephrase (filing an issue would modify remote state that isn't captured in the working tree diff, and it's not appropriate for this stage): The only gate in `package.json` is `test`. Re-running it to confirm nothing regressed: All 35 tests pass. Final verification of the working-tree change: All gates green and both decisions are made: - **Medium** — resolved by rewording Observation 3 (filing a remote issue would mutate state outside the captured working-tree diff; the reword was the finding's explicitly offered alternative). - **Low** — left unresolved: amending `0febc4f` requires rewriting git history, which this stage is prohibited from doing (no commits/stages/pushes), and reconciling AGENTS.md with the bare `<N>:` convention is a repo-owner policy call (practice is split — commits use `<N>:`, PR titles use `issue-<N>:` per PR #96). # Review Findings — 89 ## Summary Docs-only change adding a "Results — issue #89" validation record to `docs/IMPLEMENTATION_PLAN.md`. Remediation: reworded Observation 3's false claim ("Flagged for a follow-up issue" — no such issue exists) to "Should be flagged for a follow-up issue (none filed as of 2026-08-28)". The Low commit-message-format finding remains unresolved: amending `0febc4f` requires rewriting git history (prohibited in this stage), and reconciling AGENTS.md with the bare `<N>:` convention is a repo-owner policy decision. Gates re-run after the fix: `npm test` → 35/35 pass; build N/A and lint N/A (no such scripts in `package.json`; docs-only change). ## Critical (none) ## High (none) ## Medium - [x] docs/IMPLEMENTATION_PLAN.md:117 — Observation 3 asserts the README-vs-manifest victorialogs mismatch was "Flagged for a follow-up issue", but no such issue exists in the Forgejo project (searched `victorialogs` → 0 issues; `manifest`/`README` → only unrelated issues). The claim is currently false and will mislead readers into thinking tracking exists. Suggested fix: file the follow-up issue and reference its number here (e.g. "flagged as #<N>"), or reword to "should be flagged for a follow-up issue". ## Low - [ ] docs/IMPLEMENTATION_PLAN.md:96 — Commit `0febc4f` message is `89: Validate install:local from a scratch clone (fresh machine)`, but AGENTS.md requires the `issue-<N>: <summary>` format for issue-driven work (e.g. `issue-42: add user auth`). It matches this repo's recent history (commits 70–88 all use bare `<N>:`), so it is consistent with practice but deviates from the letter of the standard. Suggested fix: amend to `issue-89: Validate install:local from a scratch clone (fresh machine)` before merge, or reconcile AGENTS.md with the established `<N>:` convention.

david added 2 commits

2026-08-28 01:25:11 +00:00