Skip to content

fix: portable observer mktemp template (#2417) + correct false attribution-disabled doc claim (#2426)#2430

Open
devin-ai-integration[bot] wants to merge 1 commit into
devin/1782854552-cleanup-issue-clusterfrom
devin/1783038912-observer-mktemp-doc-fixes
Open

fix: portable observer mktemp template (#2417) + correct false attribution-disabled doc claim (#2426)#2430
devin-ai-integration[bot] wants to merge 1 commit into
devin/1782854552-cleanup-issue-clusterfrom
devin/1783038912-observer-mktemp-doc-fixes

Conversation

@devin-ai-integration

@devin-ai-integration devin-ai-integration Bot commented Jul 3, 2026

Copy link
Copy Markdown

What Changed

Two small, independent reliability/correctness fixes. Stacked on #2408 (base branch devin/1782854552-cleanup-issue-cluster) because main is currently red — see note at the bottom.

#2417 — observer wedges on macOS (malformed mktemp template)

  • skills/continuous-learning-v2/agents/observer-loop.sh: the analysis temp template was ecc-observer-analysis.XXXXXX.jsonl. BSD/macOS mktemp only substitutes a trailing run of X, so the suffix after the Xs produced a literal, non-random filename — the second cycle then failed with File exists and permanently wedged the observer. Moved the random run to the end: ecc-observer-analysis.jsonl.XXXXXX.
  • Added tests/hooks/observer-loop-mktemp.test.js: static regression that scans every mktemp "..." template in observer-loop.sh and asserts the template ends in X+ (fails the old shape, passes the fixed one).
  • Note: the issue's Findings 2 (archive-on-failure) and 3 (--max-turns 20 too low) are already fixed on main (bug: observer archives observations even when Claude analysis fails, dropping the batch from re-analysis #2370, Observer default MAX_TURNS=20 insufficient for MAX_ANALYSIS_LINES=500 #2035); only the mktemp template remained.

#2426 — false "attribution disabled globally" claim

  • The git-workflow rule asserted Attribution disabled globally via ~/.claude/settings.json, but ECC ships no such setting, so the statement is false for anyone installing the rules. Reworded it into an instruction (set "includeCoAuthoredBy": false in ~/.claude/settings.json) across every copy: rules/common/git-workflow.md, the .cursor and .kiro mirrors, and the 7 doc translations (ja-JP, es, pt-BR, zh-TW, zh-CN, tr, ko-KR). Code token and path kept verbatim; prose translated per locale.

Why This Change

Both are documentation/reliability drift: the observer silently stopped producing instincts on macOS after one interrupted cycle, and the rule pack asserted a harness state it never establishes.

Testing Done

  • Automated tests pass locally (node tests/run-all.js) — 2947 passed, 0 failed (on the stacked branch)
  • npm ci --ignore-scripts succeeds locally
  • Edge cases considered and tested

Type of Change

  • fix: Bug fix
  • docs: Documentation

Note on base branch / CI

main is currently red on the full matrix from four pre-existing infra breakages, all fixed by #2408: (1) package-lock.json eslint desync (npm ci), (2) yarn.lock desync under Yarn 4 (YN0028), (3) missing pyyaml (Python Tests ModuleNotFoundError: No module named 'yaml'), (4) Windows bash-probe in plugin-hook-bootstrap.js. Rather than duplicate all four here (they'd conflict with #2408), this PR is stacked on #2408 so its diff stays limited to the two fixes above. Once #2408 merges to main, this can be retargeted to main cleanly.

Link to Devin session: https://app.devin.ai/sessions/408d7d92c162491d98c736d2a144a96c

@devin-ai-integration devin-ai-integration Bot requested a review from affaan-m as a code owner July 3, 2026 00:58
@affaan-m affaan-m self-assigned this Jul 3, 2026
@devin-ai-integration

Copy link
Copy Markdown
Author

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

  • Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
  • Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

  • Disable automatic comment, CI, and merge conflict monitoring

@ecc-tools

ecc-tools Bot commented Jul 3, 2026

Copy link
Copy Markdown
Contributor

ECC bundle files are already tracked in this repository. Skipping generation of another bundle PR.

@greptile-apps

greptile-apps Bot commented Jul 3, 2026

Copy link
Copy Markdown
Contributor

Greptile Summary

This PR fixes observer temp-file portability and updates attribution guidance. The main changes are:

  • Moved the observer analysis mktemp random suffix to the end of the template.
  • Added a regression test for observer-loop mktemp templates.
  • Reworded git-workflow attribution notes across the canonical rule and mirrored docs.

Confidence Score: 5/5

This looks safe to merge.

  • No blocking issues found in the changed code.

Important Files Changed

Filename Overview
skills/continuous-learning-v2/agents/observer-loop.sh Updates the observer analysis temp-file template so BSD/macOS mktemp can randomize it correctly.
tests/hooks/observer-loop-mktemp.test.js Adds a test that checks quoted observer-loop mktemp templates end with the random X run.
rules/common/git-workflow.md Changes the attribution note from a global-state claim to a user setting instruction.

Reviews (3): Last reviewed commit: "fix(observer): portable mktemp template ..." | Re-trigger Greptile

Comment thread rules/common/git-workflow.md
@ecc-tools

ecc-tools Bot commented Jul 3, 2026

Copy link
Copy Markdown
Contributor

ECC bundle files are already tracked in this repository. Skipping generation of another bundle PR.

… false attribution-disabled claim in git-workflow docs (#2426)

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
@devin-ai-integration devin-ai-integration Bot changed the base branch from main to devin/1782854552-cleanup-issue-cluster July 3, 2026 01:27
@devin-ai-integration devin-ai-integration Bot force-pushed the devin/1783038912-observer-mktemp-doc-fixes branch from f2a99ec to 4d8fa0b Compare July 3, 2026 01:27
@ecc-tools

ecc-tools Bot commented Jul 3, 2026

Copy link
Copy Markdown
Contributor

ECC bundle files are already tracked in this repository. Skipping generation of another bundle PR.

1 similar comment
@ecc-tools

ecc-tools Bot commented Jul 3, 2026

Copy link
Copy Markdown
Contributor

ECC bundle files are already tracked in this repository. Skipping generation of another bundle PR.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant