AGENT SKILL

create-inspection

Investigate a specific problem and surface viable fix paths with trade-offs, saved to docs/inspections/. Triggers on: "how should I fix", "what are my options", "investigate before I touch".

License 许可
MIT
最近更新
2026-10-26
Stars
★ 0
Forks
0

01 前言 Front Matter

name
create-inspection
description
Investigate a specific problem and surface viable fix paths with trade-offs, saved to docs/inspections/. Triggers on: "how should I fix", "what are my options", "investigate before I touch".
allowed-tools
Read Grep Glob Bash Write
argument-hint
[problem]

03 SKILL.md 正文

Investigate a specific problem and produce a structured fix-path document saved to the project's docs/inspections/ directory.

Trigger: when the user asks "how should I fix X", "what are my options for Y", "investigate Z before I touch it", or triggers directly via /create-inspection. Also useful before invoking the /fix-with-evidence skill to pre-evaluate approaches.

Arguments: $ARGUMENTS — a description of the problem or subject to inspect (e.g. "auth token refresh race condition", "N+1 queries in /api/dashboard", "flaky E2E login test"). Required — if empty, ask the user.

Purpose

/create-inspection is not an audit (it doesn't enumerate all issues) and not a fix (it doesn't implement anything). Its output is a decision document: here is what is broken, here are the viable paths to fix it, here is a recommended approach with rationale. Sits between the /create-audit skill (find problems) and the /fix-with-evidence skill (implement the fix).

Steps

  1. Parse the subject from $ARGUMENTS. If empty or vague, ask the user to describe the symptom or problem.

  2. Locate the problem in the codebase. Use Grep, Glob, and Read to find the relevant files, functions, configs, and tests. Do not guess — verify every claim with a file:line citation.

  3. Diagnose root cause. Read enough code to understand why the problem occurs, not just where. If there are tests, run them to observe the failure. If the issue is behavioural (performance, flakiness), read logs or run diagnostic commands.

  4. Generate at least 2 fix options. For each option:

    • Describe the approach concisely
    • List the files/components touched
    • State the trade-offs (complexity, blast radius, test surface, reversibility)
    • Estimate effort: low (< 1 hr), medium (1–4 hrs), high (> 4 hrs)
    • Assess risk: low, medium, high
  5. Pick a recommended option and explain why it best balances correctness, risk, and effort. If options are genuinely tied, say so and flag the deciding factor for the user.

  6. Generate the inspection document with this structure:

    # Inspection: <Subject> — <YYYY-MM-DD>
    
    <One-sentence description of the problem and its impact.>
    
    ## Root Cause
    
    <Evidence-backed explanation of why the problem occurs. Include file:line citations.>
    
    ## Scope
    
    Files and components relevant to this problem.
    
    | File / Component | Role in the problem |
    | ---------------- | ------------------- |
    | ...              | ...                 |
    
    ## Fix Options
    
    ### Option 1: <Name>
    
    **Approach:** <description>
    **Files touched:** <list>
    **Effort:** low | medium | high
    **Risk:** low | medium | high
    **Trade-offs:** <pros and cons>
    
    ### Option 2: <Name>
    
    ...
    
    ## Recommendation
    
    **Use Option N: <Name>**
    
    <2-3 sentences explaining why this option is best for the current situation.
    Call out any assumptions (e.g. "assumes test coverage exists for the path").>
    
    ## Next Step
    
    Run `/fix-with-evidence <subject>` using the recommended option,
    or proceed manually using the approach above.
    
  7. Generate a filename in the format: <topic-slug>-<YYYY-MM-DD>.md (e.g. auth-refresh-race-2026-04-14.md). Use lowercase kebab-case.

  8. Write the file to docs/inspections/<filename>. Create the docs/inspections/ directory if it doesn't exist.

  9. Report to the user: show the file path, the root cause in one sentence, and the recommended option. Do not dump the full document into chat.

Rules

  • Every root-cause claim must be backed by a file:line citation or command output. No assumptions.
  • Always provide at least 2 distinct fix options. Single-option "inspections" are just instructions.
  • Never implement a fix. Surface options and recommend. The user decides.
  • Do not commit the inspection file. Leave it untracked for the user to review.
  • If the problem cannot be localized (e.g. no repro, insufficient context), state that clearly and list what additional information is needed rather than producing a speculative document.
  • Keep the document concise. Tables over prose where possible. No filler.

04 同仓库更多 Skills

kaiohenricunha
kaiohenricunha/dotbabel

yolo

Start temporary autonomy for the current task or an explicit scope with a required --for expiry. Codex permits scoped pushes, merges, deploys, and releases after strict testing and validation; Full access is a separate runtime setting. Claude Code uses a session permission hook with a protected-action denylist.

开发工程师
★ 0
kaiohenricunha
kaiohenricunha/dotbabel

veracity-audit

Audit a data pipeline for Veracity and Value. Dispatches data-scientist, compliance-auditor, and data-engineer agents with project context injected at dispatch time. All source paths are supplied via flags — no defaults, no project assumptions. Subcommands: audit (full pipeline walk), score-check (scoring math only), gate-check (gate coverage only), source-trace <label> (single source end-to-end). Invoke when: "audit pipeline", "veracity check", "scoring math correct?", "check gates", "trace source", "verify quality gates", "formula correct?".

律师与法务Skill与智能体开发
★ 0
kaiohenricunha
kaiohenricunha/dotbabel

validate-spec

Audit an already-implemented spec against the codebase. Walks each constraint (ARCH-N, PERF-N, KD-N, etc.) and acceptance criterion, grounds findings in file:line evidence, runs the spec.json acceptance_commands, and writes a single audit doc to docs/audits/. Use whenever the user asks to "validate a spec", "audit a spec", "check if spec is implemented", "verify the spec is done", "is this spec really done", or otherwise wants closure on spec-driven work. Read-only against the spec — produces an audit, never modifies the spec itself.

★ 0
kaiohenricunha
kaiohenricunha/dotbabel

terragrunt-specialist

Deep-dive Terragrunt hierarchy review, DRY pattern audits, and run-all orchestration analysis. Use for structured investigations of multi-environment Terragrunt layouts, dependency graphs, remote state config, and hook correctness. Triggers on: "Terragrunt audit", "run-all review", "dependency block", "DRY pattern review", "env hierarchy audit", "mock_outputs", "terragrunt hooks".

运维与部署
★ 0
kaiohenricunha
kaiohenricunha/dotbabel

terraform-specialist

Deep-dive Terraform architecture review, module design, state management, and migration. Use for structured investigations of Terraform workspaces, provider configuration, module coupling, import workflows, and test coverage. Triggers on: "Terraform audit", "module review", "state management", "Terraform import", "workspace design", "provider config review", "Terraform testing".

测试与质量运维与部署
★ 0
kaiohenricunha
kaiohenricunha/dotbabel

spec

Create structured engineering specs through interactive pairing. Use when the user wants to create a spec, design doc, technical specification, architecture document, RFC, or engineering plan. Triggers on "let's spec this out", "create a spec", "design doc", "write a technical plan", "plan the architecture". Works for greenfield and brownfield projects. Outputs to docs/specs/.

设计师视觉设计
★ 0