ワンクリックで
solve-issue-local
Repo-specific build, test, and PR instructions for mage-bench issue solving.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Repo-specific build, test, and PR instructions for mage-bench issue solving.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
Debug a failing golden test from GitHub Actions logs and fix the root cause without papering over nondeterminism.
Investigate a game from raw logs, trace bugs to source code, and file detailed issues. Use for deeper debugging than export-only analysis.
Repo-specific validation and PR instructions for mage-bench.
Find Magic cards for compact golden-test scenarios using existing repo goldens, curated references, and Scryfall search recipes.
Analyze exported game files to assess gameplay, model decisions, and likely bugs without raw logs. Use for quick game triage.
Write or update low-noise golden prompt tests in tests/, including minimal decks, replay scripts, and golden regeneration.
SOC 職業分類に基づく
| name | solve-issue-local |
| description | Repo-specific build, test, and PR instructions for mage-bench issue solving. |
This file provides repo-specific instructions for the global /solve-issue skill.
make check # Full lint + typecheck + tests (the one command to rule them all)
make build # Java build only
make test # Python tests only
make lint # Python lint only
make check VERBOSE=1 over launching a second blind make check.Mage.Client.Bridge/, use a module-scoped Maven loop for fast feedback while iterating (e.g. mvn -pl Mage.Client.Bridge -DskipTests compile). Still finish with the full make check before finalizing.After implementation, run /simplify to review the changed code for reuse, quality, and efficiency. Fix any issues found.
Also inspect website/package-lock.json before committing. make check / website tooling can add incidental "peer": true lockfile churn even when you did not intentionally change website dependencies; drop unrelated lockfile noise so the issue PR stays scoped.
If you run targeted Ruff fixes or formatting on "changed files only", remember that git diff --name-only omits newly added untracked files. Include untracked paths from git status too, or lint can still fail on the brand-new files you just created.
conftest.py and shared test fixtures for hardcoded repo paths before relying on the full test suite.patch("old.module.symbol"); they do not show up in import grep but will fail at runtime once the old module disappears.puppeteer/ into top-level tests/), also update repo plumbing that discovers or consumes them: Makefile lint/test targets, src/magebench/cli/checks/quiet_check.py triggers, pytest config location, and any website/utilities that read golden assets by path. The file move alone is not enough to keep make check green.If MCP tools, pilot logic, or priority semantics change enough to make game results non-comparable, bump HARNESS_EPOCH in src/magebench/game/harness_epoch.py.
- [ ] Implement the changes
- [ ] Add/update tests
- [ ] Run `make check` (lint, typecheck, tests)
- [ ] Delete the issue file and include deletion in the commit
- [ ] Run `/simplify` to review changed code
- [ ] Push final changes: `git push origin HEAD`
- [ ] Finalize PR: `issue-finalize-pr --title "..." --body "..."`
- [ ] Watch CI: `issue-watch-pr`