Agent skill
check-updates
Check upstream repo changes since last analysis, assess impact on docs and demos, or stamp current commit
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/check-updates
SKILL.md
Check Updates
Check if upstream repos have changed since the last analysis, assess the impact on docs and demos, or stamp the current commit as analyzed.
Trigger
/check-updates [--stamp] [agent-name]
- Without arguments: check all agents with
analyzed_commit - With
agent-name: check only that agent - With
--stamp agent-name: update commit tracking to current submodule HEAD
Mode 1: Check (default)
/check-updates [agent-name]
Steps
-
Read agents.yaml — find agents with
analyzed_commitfield. Ifagent-nameis given, filter to that one. -
Get remote HEAD — for each agent, extract
{owner}/{repo}from therepo:URL and run:bashgh api repos/{owner}/{repo}/commits?per_page=1 --jq '.[0].sha' -
Compare — if
analyzed_commit== remote HEAD → report "UP TO DATE". Otherwise:bashgh api repos/{owner}/{repo}/compare/{analyzed_commit}...{remote_head} -
Classify changed files by dimension impact:
File pattern Likely affected dimension **/main.*,**/cli.*,**/core/**D1 Architecture, D2 Agent Loop **/tool*,**/command*,**/action*D3 Tool System **/prompt*,**/system*,**/template*D4 Prompt Engineering **/context*,**/token*,**/repomap*D5 Context Management **/error*,**/retry*,**/recover*D6 Error Handling README*,CHANGELOG*D1 Overview **/test*,*.md,.github/**Low impact -
Assess demo impact — read
demos/<agent>/README.mdto find completed demos ([x]). For each demo:- Read
demos/<agent>/<demo>/README.mdfor source file references (look for lines like核心源码:, code file paths in "原理" or "相关文档" sections) - Match referenced files against changed files in the compare diff
- If no explicit reference, use fuzzy match: demo name ↔ changed file name
- Report which demos may need updating
- Read
-
Output structured report:
## <agent-name>
**Status**: ⚠️ DRIFT DETECTED (N commits behind)
**Analyzed at**: <commit_short> (<date>)
**Upstream HEAD**: <remote_short> (<date>)
**Compare**: <github_compare_url>
### 文档影响
- **高影响维度**: D2, D3(核心文件变更)
- **低影响维度**: D1(仅文档变更)
- **无影响维度**: D4, D5, D6, D7, D8
### Demo 影响
- ⚠️ **repomap**: `aider/repomap.py` 有 +30/-10 行变更 → 可能需要更新 demo
- ✅ **search-replace**: 相关文件无变更
- ✅ **reflection**: 相关文件无变更
- ⚠️ **architect**: `aider/coders/architect_coder.py` 有变更 → 可能需要更新 demo
### 建议
- [ ] 更新 submodule: `npm run update -- <agent>`
- [ ] 重新审查 D2, D3 部分
- [ ] 检查 repomap, architect demo 是否需要更新
- [ ] 或运行 `/check-updates --stamp <agent>` 确认当前内容仍有效
For agents that are up to date:
## <agent-name>
**Status**: ✅ UP TO DATE
**Analyzed at**: <commit_short> (<date>)
Mode 2: Stamp
/check-updates --stamp <agent-name>
Update commit tracking after reviewing changes or completing a re-analysis.
Steps
-
Read current HEAD from
projects/<agent-name>/:bashgit -C projects/<agent-name> rev-parse HEADIf the submodule doesn't exist, abort with error.
-
Update agents.yaml — set
analyzed_commitandanalyzed_date(today's date) for the agent. -
Update doc header — in
docs/<agent-name>.md, update the> Analyzed at commit:line with the new SHA and date. If the line doesn't exist, add it after the> Repo:line. -
Update demo overview — in
demos/<agent-name>/README.md, update the> Based on commit:line with the new SHA and date. If the line doesn't exist, add it after the first description paragraph. -
Report what was updated.
Key Design Decisions
- Uses
gh apiinstead of localgit logto avoid shallow clone limitations - Check mode is read-only — it does NOT update the submodule or any files
- File→dimension mapping is assessed by Claude using the pattern table above as guidance
- Demo impact assessment uses source file references in demo READMEs + fuzzy file name matching
MVP Coverage Hint
If the agent's demo overview (demos/<agent>/README.md) uses the three-tier format (MVP 组件/进阶机制/完整串联), append a line at the end of the report:
> 💡 如需检查 MVP 覆盖缺口,运行 `/audit-coverage <agent>`
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?