Quick Review Skill

Lightweight wrapper that triggers Claude Octopus grasp → tangle workflow for fast, consensus-driven code review.

When This Skill Activates

Auto-invokes when user says:

• "review this code"
• "check this PR"
• "quality check the implementation"
• "review my changes"
• "what's wrong with this code"

What It Does

Two-Phase Workflow:

1. Grasp (Define): Multi-agent consensus on what needs review

   • All agents independently identify issues
   • Synthesize into consensus list of concerns
   • Prioritize by severity and impact

2. Tangle (Develop): Parallel implementation review

   • Each agent reviews specific aspects (security, performance, maintainability)
   • Quality gate ensures ≥75% agreement
   • Synthesized findings with actionable recommendations

Usage

User: "Review this authentication module for security issues"

Claude: *Activates octopus-quick-review skill*
        *Runs: ${CLAUDE_PLUGIN_ROOT}/scripts/orchestrate.sh grasp "Review authentication module"*
        *Then: ${CLAUDE_PLUGIN_ROOT}/scripts/orchestrate.sh tangle "Implement review findings"*

Implementation

When this skill is invoked, Claude should:

1. Detect intent: User wants code review
2. Invoke grasp: Get multi-agent consensus on review scope

   ${CLAUDE_PLUGIN_ROOT}/scripts/orchestrate.sh grasp "[user's review request]"

3. Invoke tangle: Parallel review with quality gates

   ${CLAUDE_PLUGIN_ROOT}/scripts/orchestrate.sh tangle "[synthesized review scope]"

4. Present findings: Format results for user

Output Format

## Review Summary

**Consensus Issues** (from grasp phase):
1. Authentication bypass in login handler
2. Missing rate limiting
3. Weak password validation

**Detailed Analysis** (from tangle phase):

### Security (Agent: Codex)
- Critical: SQL injection risk in user input
- Medium: Session tokens not rotated

### Performance (Agent: Gemini)
- High: N+1 queries in user lookup
- Low: Inefficient password hashing

**Quality Gate**: PASSED (85% agreement)

**Recommendations**:
1. Implement parameterized queries
2. Add rate limiting middleware
3. Use bcrypt with cost factor 12

Why Use This Instead of Full Embrace?

Aspect	Quick Review	Full Embrace
Speed	2-5 min	5-10 min
Phases	2 (grasp, tangle)	4 (probe, grasp, tangle, ink)
Best For	Code review, PR checks	New features, architecture
Depth	Focused consensus	Deep research + validation

Configuration

Respects all octopus configuration:

• --autonomy: Control quality gate behavior
• --quality: Set consensus threshold (default 75%)
• --provider: Force specific AI provider
• --cost-first: Optimize for speed/cost

Example Scenarios

Scenario 1: PR Review

User: "Review PR #123 for security issues"
→ Grasp: Identify all security concerns
→ Tangle: Detailed OWASP Top 10 analysis
→ Output: Prioritized security findings

Scenario 2: Performance Check

User: "Why is my API so slow?"
→ Grasp: Consensus on performance bottlenecks
→ Tangle: Parallel analysis (DB, network, compute)
→ Output: Optimization recommendations

Scenario 3: Quick Sanity Check

User: "Does this look okay?"
→ Grasp: Quick consensus (basic issues only)
→ Tangle: Fast pass with --cost-first
→ Output: Green light or red flags

Related Skills

• octopus-security (squeeze workflow) - For adversarial security testing
• octopus-research (probe workflow) - For deep investigation
• Full embrace - For complete Double Diamond workflow

Technical Notes

• Uses existing grasp/tangle commands from orchestrate.sh
• No new code required - pure workflow coordination
• Leverages quality gates for reliability
• Session recovery supported (can resume if interrupted)