a-team
AGENTS.md24 starsChanged 24 days agoMITView on GitHub
Model values below are Claude Code defaults (model: frontmatter). On other platforms (Codex, Cursor, OpenCode, GitHub Copilot CLI), select the equivalent tier in your platform's model settings: Tier 1 = o3/GPT-4o/Gemini 2.5 Pro · Tier 2 = GPT-4o/Gemini Flash · Tier 3 = GPT-4o-mini/Gemini Flash-Lite.
In this repository22 other files it gives its agents.
- .claude/rules/agents.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- api-contract-firstSkill
- architecture-auditSkill
- brainstormingSkill
- data-migrationSkill
- dispatching-parallel-agentsSkill
- executing-plansSkill
- finishing-a-development-branchSkill
- five-whysSkill
- incident-responseSkill
- performance-auditSkill
- skill-duplication-auditSkill
- smart-initSkill
- subagent-driven-developmentSkill
- systematic-debuggingSkill
- test-driven-developmentSkill
- using-a-teamSkill
- using-git-worktreesSkill
- verification-before-completionSkill
- writing-plansSkill
- writing-skillsSkill
The file
Files
A Team — Agent & Skill Roster v1.4.1
Model values below are Claude Code defaults (
model:frontmatter). On other platforms (Codex, Cursor, OpenCode, GitHub Copilot CLI), select the equivalent tier in your platform's model settings: Tier 1 = o3/GPT-4o/Gemini 2.5 Pro · Tier 2 = GPT-4o/Gemini Flash · Tier 3 = GPT-4o-mini/Gemini Flash-Lite.
Core Engineering Agents
| Agent | Tier | Trigger |
| orchestrator | Tier 1 (claude-opus-4-8) | /orchestrate init|morning|tick|report |
| architect | Tier 1 (claude-opus-4-8) | New feature design, tech decisions |
| planner | Tier 2 (claude-sonnet-4-6) | Implementation plan before coding |
| code-reviewer | Tier 2 (claude-sonnet-4-6) | After every code change (mandatory) |
| security-reviewer | Tier 2 (claude-sonnet-4-6) | Auth, API, input, DB code |
| tdd-guide | Tier 2 (claude-sonnet-4-6) | New features, bug fixes |
| refactor-cleaner | Tier 2 (claude-sonnet-4-6) | Dead code, duplication cleanup |
| build-error-resolver | Tier 2 (claude-sonnet-4-6) | Build or type errors |
| e2e-runner | Tier 2 (claude-sonnet-4-6) | Critical user journeys |
| doc-updater | Tier 3 (claude-haiku-4-5) | Post-feature docs & codemaps |
| debugger | Tier 2 (claude-sonnet-4-6) | Any bug (root-cause first) |
Language & Platform Specialists
| Agent | Language / Platform | Trigger |
| database-reviewer | PostgreSQL | SQL, migrations, schema |
| go-reviewer | Go | All .go changes |
| python-reviewer | Python | All .py changes |
| rust-reviewer | Rust | All .rs changes |
| kotlin-reviewer | Kotlin / Android | All .kt changes |
| swift-reviewer | Swift / iOS | All .swift changes |
| flutter-reviewer | Flutter / Dart | All .dart changes |
| typescript-reviewer | TypeScript / React / Frontend | All .ts/.tsx changes |
Operational Agents
| Agent | Tier | Purpose |
| chief-of-staff | Tier 2 (claude-sonnet-4-6) | Email/Slack/Calendar triage |
| loop-operator | Tier 2 (claude-sonnet-4-6) | Autonomous loop safety |
| harness-optimizer | Tier 3 (claude-haiku-4-5) | Pipeline auditor — detects rule evasion, blocks merge |
| infra-reviewer | Tier 2 (claude-sonnet-4-6) | Terraform, Docker, K8s, CI/CD — IaC changes |
| compliance-reviewer | Tier 2 (claude-sonnet-4-6) | GDPR, COPPA, PCI-DSS, SOC2, HIPAA |
| ai-reviewer | Tier 2 (claude-sonnet-4-6) | LLM code — prompt injection, token cost, tool call safety |
| performance-profiler | Tier 3 (claude-haiku-4-5) | Systematic profiling — baseline, bottleneck, validate |
Skill Library
| Skill | Gate | Purpose |
| using-a-team | Hard | Meta-skill injected every session — enforces all others |
| verification-before-completion | Hard | Prove it works before claiming done |
| brainstorming | Hard | Design before code — always |
| test-driven-development | Hard | RED first, always |
| systematic-debugging | Hard | Root cause before fix |
| subagent-driven-development | Workflow | Execute plans with fresh subagents + two-stage review |
| using-git-worktrees | Workflow | Isolated workspace before development |
| executing-plans | Workflow | Run a written plan in current session |
| dispatching-parallel-agents | Workflow | 2+ independent problems concurrently |
| writing-plans | Workflow | Create structured implementation plans |
| finishing-a-development-branch | Workflow | Wrap up a branch before merge |
| writing-skills | Meta | Create new A Team skills via TDD |
| five-whys | Workflow | Structured causal analysis |
| smart-init | Workflow | ROADMAP-aware initialization |
| skill-duplication-audit | Workflow | Audit overlapping skills across packs |
| architecture-audit | Workflow | Evidence-backed architecture review |
| api-contract-first | Hard | Design OpenAPI/protobuf contract before any endpoint |
| incident-response | Workflow | Production incident — detect, contain, diagnose, resolve, post-mortem |
| data-migration | Hard | Safe schema migrations with rollback strategy |
| performance-audit | Workflow | Baseline → profile → single change → measure again |
Command Reference
/orchestrate init → Read INIT.md, prune team, generate routing
/orchestrate morning → Plan today's tasks
/orchestrate tick → Process completed results
/orchestrate report → End-of-day telemetry
/plan → Create implementation plan
/code-review → Review staged/recent changes
/security-review → Security audit
/refactor → Find & remove dead code
/build-fix → Fix build errors (minimal diffs)
/debug → Systematic root-cause debugging
/e2e → Run end-to-end test suite
/quality-gate → Full pre-merge check (code + security)
/feature → Full feature workflow (brainstorm→plan→implement→review)Mandatory Skill Invocation Map
| You are about to... | Required first |
| Write a new feature | brainstorming |
| Start dev work | using-git-worktrees |
| Write any code | test-driven-development |
| Fix any bug | debugger agent |
| Claim anything is "done" | verification-before-completion |
| Review code | code-reviewer agent |
| Merge a branch | harness-optimizer audit → /quality-gate |
Parallel Dispatch Patterns
For independent problems, dispatch in parallel:
# 3 failing test files → 3 parallel debugger agents
# Security + code review → security-reviewer + code-reviewer in parallel
# Go + Python changes in same PR → go-reviewer + python-reviewer in parallelCross-Platform Support
| Platform | Config |
| Claude Code | .claude/ (native) |
| Cursor | .cursor-plugin/plugin.json |
| Codex | .codex-plugin/plugin.json |
| GitHub Copilot CLI | .copilot-plugin/plugin.json |
| OpenCode | .opencode/commands/ |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.