goal
CommandSet, check, or complete the persistent project goal stored in .goal.md. Trigger as /coco-goal or /coco:goal.
@coco-researchInstall CommandDevelopment
Command
/review
Use when the /team router routes action=review, or the user asks for existing code, docs, or a diff reviewed. Scope mapping, 3-5 specialist lenses, cross-review, then a principal synthesis of severity-ranked findings.
$ npx -y skills add coco-research/coco --agent claude-code
Command definition
review.mddescription: "Use when the /team router routes action=review, or the user asks for existing code, docs, or a diff reviewed. Scope mapping, 3-5 specialist lenses, cross-review, then a principal synthesis of severity-ranked findings."
/team review — Review Pipeline
> Called by team.md router when action is `review`.
Role Selection Bias
| Layer | Preferred Roles | Count | |-------|----------------|-------| | L1 | technical-analyst, security-analyst | 2 | | L2 | (domain-dependent engineers) + senior-pm | 3-5 | | L3 | domain-accuracy, doc-quality, grammar-editor, accessibility-specialist (if frontend domain) | 2 | | L4 | principal-architect | 1 |
Key Difference from Old System
**Old:** 5 identical reviewer clones with the same prompt → redundant findings. **New:** 3-5 specialists each reviewing through a different lens:
- Backend engineer reviews API design and data modeling
- Security analyst reviews auth flows and secrets
- Frontend engineer reviews component patterns and a11y
- PM reviews requirements coverage and user stories
- QA reviews test adequacy and edge cases
Pipeline Customization
Layer 1: Scope Mapping
L1 agents determine:
- What files/modules are in scope for review
- What changed since last review (git log analysis)
- What the original requirements/spec was
- What risk areas to focus on
Layer 2: Specialist Review
- **Mode:** `default` (read-only)
- Each L2 agent reviews through their specialist lens
- Every finding classified as: CRITICAL | MAJOR | MINOR | SUGGESTION
- Every finding includes: file path, line number, specific quote, suggested fix
**If GSD active:** Map findings back to requirements from REQUIREMENTS.md.
**Toolkit integration:**
- Check team:toolkit.md for "Code Review" entry for review methodology
- If GSD active, use GSD phase context for requirements traceability
Layer 3: Cross-Review
L3 agents check the reviewers' work:
- Are findings accurate? (domain-accuracy verifies claims)
- Are severity classifications consistent?
- Any overlapping or contradictory findings?
Layer 4: Synthesis
Principal produces:
- Prioritized findings list (deduplicated, severity-consistent)
- Architectural assessment (is the overall approach sound?)
- Recommended action items with effort estimates
Read more
description: "Use when the /team router routes action=review, or the user asks for existing code, docs, or a diff reviewed. Scope mapping, 3-5 specialist lenses, cross-review, then a principal synthesis of severity-ranked findings."
/team review — Review Pipeline
> Called by team.md router when action is `review`.
Role Selection Bias
| Layer | Preferred Roles | Count | |-------|----------------|-------| | L1 | technical-analyst, security-analyst | 2 | | L2 | (domain-dependent engineers) + senior-pm | 3-5 | | L3 | domain-accuracy, doc-quality, grammar-editor, accessibility-specialist (if frontend domain) | 2 | | L4 | principal-architect | 1 |
Key Difference from Old System
**Old:** 5 identical reviewer clones with the same prompt → redundant findings. **New:** 3-5 specialists each reviewing through a different lens:
- Backend engineer reviews API design and data modeling
- Security analyst reviews auth flows and secrets
- Frontend engineer reviews component patterns and a11y
- PM reviews requirements coverage and user stories
- QA reviews test adequacy and edge cases
Pipeline Customization
Layer 1: Scope Mapping
L1 agents determine:
- What files/modules are in scope for review
- What changed since last review (git log analysis)
- What the original requirements/spec was
- What risk areas to focus on
Layer 2: Specialist Review
- **Mode:** `default` (read-only)
- Each L2 agent reviews through their specialist lens
- Every finding classified as: CRITICAL | MAJOR | MINOR | SUGGESTION
- Every finding includes: file path, line number, specific quote, suggested fix
**If GSD active:** Map findings back to requirements from REQUIREMENTS.md.
**Toolkit integration:**
- Check team:toolkit.md for "Code Review" entry for review methodology
- If GSD active, use GSD phase context for requirements traceability
Layer 3: Cross-Review
L3 agents check the reviewers' work:
- Are findings accurate? (domain-accuracy verifies claims)
- Are severity classifications consistent?
- Any overlapping or contradictory findings?
Layer 4: Synthesis
Principal produces:
- Prioritized findings list (deduplicated, severity-consistent)
- Architectural assessment (is the overall approach sound?)
- Recommended action items with effort estimates
Ships withcoco
495 personas across 13 departments, 229 skills[^installed] and 388 commands in the full catalog (coco 1.9.0, 2026-10-21). Agent skills, slash commands and subagents for Claude Code, Cursor, Codex and other AI coding agents.
796200 skills53 agents
Stars on GitHub
Momentum
- stars in total
- 796
Other commands on coco.
@coco-researchInstall Commandmermaid
CommandUse when the user asks for a diagram or chart: flowchart, sequence, state, class, ER, or XY chart, as SVG, HTML, or terminal…
@coco-researchInstall Commandread
CommandShow latest emails from a specific person. Usage: /email-read alice
@coco-researchInstall Commandreply
CommandDraft a reply to a specific email. Usage: /email-reply Project Phase 2 Contract Data
@coco-researchInstall Commandsave
CommandSave matching emails to a project folder for sync processing. Usage: /email-save alice to emails/
@coco-researchInstall Commandsearch
CommandSearch emails by subject keywords. Usage: /email-search github webhook
@coco-researchInstall CommandMore commands from other plugins.
Auto-invoked
Auto-invoked
build
CommandImplement tasks incrementally — build, test, verify, commit. Add "auto" to run the whole plan in one approved pass.
@addyosmaniInstall Command