AgentSkillsCN

code-council

利用 code-router 实现多后端并行代码审查。同时运行 2–3 名 AI 审查员，使用相同的全面审查提示，随后由主审智能体核实审查结果，并综合形成最终报告。

SKILL.md
--- frontmatter
name: code-council
description: Multi-backend parallel code review using code-router. Runs 2-3 AI reviewers simultaneously with the same comprehensive review prompt, then the host agent verifies findings and synthesizes a final report.
compatibility: Requires code-router skill with at least 2 backends configured (codex, claude, gemini).

code-council: Multi-Backend Code Review

Overview

Leverage code-router's multi-backend parallel execution to run multiple AI reviewers simultaneously, each performing an independent comprehensive code review. The host agent then verifies each finding and synthesizes the final report — ensuring only real issues are presented to the user.

When to Use

  • •User explicitly requests code review, architecture audit, or quality check
  • •PR review or pre-merge quality gate
  • •Module-level or file-level review
  • •A skill or command definition declares a dependency on this skill

Workflow

Step 1: Determine Review Scope

Resolve target files from user input:

  • •Explicit references: user provides file/directory paths → use @ syntax
  • •Git diff: user mentions PR, diff, or recent changes → capture diff output, pass as context
  • •Module: user names a package or directory → glob relevant source files

If scope is ambiguous, ask the user to clarify.

Step 2: Select Backends

Ask the user which backends to enable for parallel review:

Options:

  • •codex (recommended)
  • •claude (recommended)
  • •gemini (recommended)

Multi-select: Allow user to pick one or more backends

Default: All three enabled

Fallback (if AskUserQuestion is not available in your environment):

  • •Stop and ask the user explicitly: "Which backends do you want to use for review? Options: codex, claude, gemini. Default: all three."
  • •Wait for user response before proceeding

Step 3: Build & Execute Parallel Review

Construct a single code-router --parallel invocation. All backends run the same comprehensive review prompt.

Template (3 backends):

bash
code-router --parallel --backend codex <<'EOF'
---TASK---
id: review_codex
backend: codex
---CONTENT---
You are a code reviewer. Perform a comprehensive review of the target code.

Focus on:
1. Security vulnerabilities (injection, auth bypass, data exposure)
2. Logic errors and bugs
3. Performance issues
4. Error handling completeness
5. Code quality and maintainability
6. Potential edge cases

Target: [@ file references]

Output format:
## Code Review Findings
For each finding:
- **[CRITICAL|WARNING|INFO]**: one-line summary
  - Location: file:line
  - Detail: what is wrong and why it matters
  - Suggestion: concrete fix

---TASK---
id: review_claude
backend: claude
---CONTENT---
You are a code reviewer. Perform a comprehensive review of the target code.

Focus on:
1. Security vulnerabilities (injection, auth bypass, data exposure)
2. Logic errors and bugs
3. Performance issues
4. Error handling completeness
5. Code quality and maintainability
6. Potential edge cases

Target: [@ file references]

Output format:
## Code Review Findings
For each finding:
- **[CRITICAL|WARNING|INFO]**: one-line summary
  - Location: file:line
  - Detail: what is wrong and why it matters
  - Suggestion: concrete fix

---TASK---
id: review_gemini
backend: gemini
---CONTENT---
You are a code reviewer. Perform a comprehensive review of the target code.

Focus on:
1. Security vulnerabilities (injection, auth bypass, data exposure)
2. Logic errors and bugs
3. Performance issues
4. Error handling completeness
5. Code quality and maintainability
6. Potential edge cases

Target: [@ file references]

Output format:
## Code Review Findings
For each finding:
- **[CRITICAL|WARNING|INFO]**: one-line summary
  - Location: file:line
  - Detail: what is wrong and why it matters
  - Suggestion: concrete fix
EOF

Template (2 backends): Omit one ---TASK--- block

Template (1 backend): Use single ---TASK--- block

Step 4: Host Agent Verification & Synthesis (MANDATORY — DO NOT SKIP)

After code-router returns, the host agent MUST verify findings and synthesize the final report.

Note: each code-router task is a full AI agent with complete filesystem access — it can read any file, run git commands, and explore the entire project. Tasks are NOT "blind" to the project. What they lack is the user's conversation context — the host agent knows what the user has been discussing, their priorities, and their intent.

Why this step exists:

  • •User context: The host agent carries the ongoing conversation — it knows what the user cares about, prior discussions, and implicit priorities
  • •Verification: Backends may produce false positives or miss real issues — the host agent must verify each finding
  • •Interactive ability: The host agent can ask the user follow-up questions; code-router tasks are fire-and-forget
  • •Quality gate: Raw outputs should be verified and synthesized, not dumped as-is

What the host agent does:

  1. •Read all review outputs from each backend
  2. •Deduplicate: merge findings that point to the same issue, note which backends agree
  3. •Verify each finding:
    • •Re-examine the code to confirm the issue is real
    • •If a finding is false positive, mark it and explain why
    • •If backends missed something, add your own findings
  4. •Rank by severity: CRITICAL first, then WARNING, then INFO
  5. •Group by file
  6. •Output final report titled "## Code Council — Verified Review"
    • •Include a summary: total counts by severity
    • •Note which backends reported each finding

Step 5: Offer Follow-up Actions

After presenting, offer concrete next steps:

  • •Fix critical issues: generate fix tasks and run them through code-router
  • •Review more files: restart the workflow on another target
  • •Save report: write to a file if the user wants a record

Severity Definitions

  • •CRITICAL: Will cause bugs, security vulnerabilities, data loss, or crashes in production
  • •WARNING: Code smell, maintainability risk, or potential future bug under reasonable conditions
  • •INFO: Style suggestion, minor improvement, or optimization opportunity

Edge Cases

Single file, trivial size (<30 lines):

  • •Still run the full council — small code can have critical issues
  • •Reviewers will naturally produce fewer findings

User provides a git diff instead of files:

  • •Capture the diff content, pass it inline in the ---CONTENT--- sections instead of @ file references
  • •Adjust reviewer prompts to focus on "changes" rather than "code"

Critical Rules

  1. •MUST perform Step 4 (verification & synthesis) — this is the core value proposition
  2. •MUST use code-router skill — this skill depends on code-router's --parallel execution
  3. •MUST be called by host agent only — this skill is NOT designed for sub-agent use
  4. •NEVER modify source code in this skill — code-council is read-only; fixes go through a separate action

Example Invocations

Review specific files:

code
Review @src/auth/login.ts and @src/auth/session.ts using code-council

Review recent changes (PR-style):

code
Run code-council on the uncommitted changes

Audit a module:

code
code-council audit @src/payments/

Use specific backends:

code
Run code-council on @src/core/ using only codex and claude