e-code-review
Code review at one of five effort levels that match the built-in /code-review recipes step for step: low, medium, high, xhigh, max. Triggers: /e-code-review with an optional level and an optional --fix.
- License 许可
- MIT
- 最近更新
- 2026-11-15
- Stars
- ★ 7
- Forks
- 1
01 前言 Front Matter
- name
- e-code-review
- description
- Code review at one of five effort levels that match the built-in /code-review recipes step for step: low, medium, high, xhigh, max. Triggers: /e-code-review with an optional level and an optional --fix.
02 资源文件 6
- 参考文档reference/fix.md570 B
- 参考文档reference/high.md6.4 KB
- 参考文档reference/low.md2.0 KB
- 参考文档reference/max.md8.6 KB
- 参考文档reference/medium.md1.5 KB
- 参考文档reference/xhigh.md7.9 KB
03 SKILL.md 正文
e-code-review
Pick a level, load its file, run it. Each file has the same steps, angles, caps, and fields as the built-in /code-review recipe for that level. Only the wording differs, where this repo's prose rules require it. A question about /code-review asks about the built-in recipe, so answer it from the built-in text and use these files only for /e-code-review.
| Level | File | Shape |
|---|---|---|
low |
reference/low.md | 1 diff pass, no verify, at most 4 findings |
medium |
reference/medium.md | 1 careful diff pass, at most 15 findings |
high |
reference/high.md | 8 inline angles, dedup with no verify, at most 10 findings |
xhigh |
reference/xhigh.md | 10 inline angles, dedup, gap sweep, at most 15 findings |
max |
reference/max.md | 10 subagent angles, 1-vote verify, gap sweep, at most 15 findings |
Level
With no level, or an unknown one, run high.
The level sets review depth. Child agents take their model and effort from Subagent model and effort.
Target
A PR number, branch name, or file path after the level is the review target. With no target, the review reads the current diff.
Report
Each level file reports through the ReportFindings tool. When the host has no ReportFindings tool, print the findings in its place, most severe first, one per line as file:line — summary. --fix reports its outcomes the same way, with the outcome at the end of each line.
--fix
--fix works with every level. It lives in one file, reference/fix.md. No level file carries its own fix steps.
The process
Native GitHub evidence pilot
Keep one coordinator responsible for native review requests. Record the request comment and full candidate head. Read raw GitHub review records to retain commit_id; normalized connector output can omit it. Completed reviews can contain findings. The quota_notice_ids field records historical usage-limit comments. Driver replies and resolved threads provide no clean completion proof.
Use process-local GITHUB_TOKEN or GH_TOKEN authentication without printing credentials. A code-review usage-limit response stops further triggers until capacity is verified. Preserve the hold; review permission changes and credit purchases require separate authorization. Capture API bytes as UTF-8 before decoding them so Windows shell output encoding preserves review headings.
Inspect a candidate from the shared source checkout:
python packages/claude-dev-env/scripts/codex_review_observer.py <owner/repository> <pull-request> <full-head>
python -m pytest packages/claude-dev-env/scripts/tests/test_codex_review_observer.py packages/claude-dev-env/scripts/tests/test_codex_review_guidance.py -q --override-ini="addopts=" -p no:cacheprovider
The current observer holds admission for every supported diagnostic. Require captured successful native completion with full-head and attempt coverage before adding a clean verdict. After a repair push, review the new head and preserve all existing checks. Keep current repository contracts authoritative over copied hosted policy.
- Read the level and the optional
--fixflag. Load the level file. - Run it end to end, ending in its ReportFindings call.
- With
--fix, loadreference/fix.mdand run it on those findings.
04 同仓库更多 Skills
update
Fast-forwards a local repository's main branch to a remote's main after confirming the repo path and source remote through AskUserQuestion. Use when the user says '/update', 'update main', 'fast-forward main', 'sync main from origin', 'pull latest main into <path>', or 'bring main up to date'.
source-command-sr-loop
Converging cleanup loop: repeat /simplify until clean, then repeat /code-review --fix until clean
skill-builder
Skill lifecycle: classify, scaffold, write via the skill-writer-agent, self-audit, compose sub-skills, polish description triggers, enforce deterministic scripts. Triggers: build a skill, new skill workflow, improve this skill, optimize skill description, skill development lifecycle, skill modularity, skill composition, description trigger catalog, deterministic skill scripts.
session-tidy
Audit, clean, and consolidate session logs in the Obsidian vault. Fixes format drift, resolves orphaned next-steps, updates stale statuses, and generates project rollup summaries. Use when session logs feel cluttered or before starting a new project milestone. Triggers on '/session-tidy', 'tidy sessions', 'clean up session logs', 'session audit'.
session-log
Logs a session report as an HTML page, tracks vault context, extracts decisions, and outputs a /rename command. Use for /session-log, 'journal this session', 'log this work', 'session report', 'summarize/log/record this session', 'save session', 'capture session', or 'document what we did'.
run-claude-dev-env
Build, install, run, and test the claude-dev-env installer. Use when asked to run claude-dev-env, install the Claude Code config into ~/.claude, drive the installer, smoke-test it against a sandbox, or run its test suite.