AGENT SKILL
trace
Follow a request end-to-end through the stack from route to database
- License 许可
- MIT
- 最近更新
- 2026-11-10
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- trace
- type
- review
- version
- v0.1
- description
- Follow a request end-to-end through the stack from route to database
02 资源文件 2
- 参考文档agents/openai.yaml226 B
- 参考文档CHANGELOG.md311 B
03 SKILL.md 正文
Trace
Invoke as $trace.
Use this skill when the user wants to understand how a specific route, endpoint, or feature flows through the application stack.
Process
- Identify the entry point from the user's description.
- Read CLAUDE.md for architecture context.
- Trace forward through each layer: entry point → middleware → business logic → data access → response.
- Map the data flow and types at each boundary.
- Identify concerns: N+1 queries, auth gaps, type mismatches, performance issues.
Output
- Visual trace: entry → middleware → handler → service → DB → response with file:line refs
- Layer-by-layer detail: file, inputs, outputs, what it does
- Data flow: how data transforms through the stack
- Concerns found: issues identified during the trace
Constraints
- Follow actual code paths — do not guess from naming.
- Read each file in the chain.
- Do not modify any code — read-only analysis.
Default Shipping Contract
Follow the shared shipping contract convention in CLAUDE.md.
04 同仓库更多 Skills
source-checkout-allowed
Valid source-checkout-only fixture.
★ 1
deck-install-valid
Valid deck install fixture.
★ 1
codex-pack-guard-valid
Valid Codex Pack Availability Guard fixture.
★ 1
codex-missing-skill-valid
Valid Codex missing skill fallback fixture.
★ 1
codex-base-skill-fallback-valid
Valid Codex base skill fallback fixture.
★ 1
claude-pack-guard-valid
Valid Claude Pack Availability Guard fixture.
★ 1