From cdt
Multi-agent development workflow with five modes: plan (architect + PM debate), dev (developer + code-tester + QA + reviewer), full (plan with approval gate), auto (plan without gate), and bugfix (TDD triad).
How this skill is triggered — by the user, by Claude, or both
Slash command
/cdt:cdtThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Multi-agent development workflow with five modes. Pick one based on the user's needs:
Multi-agent development workflow with five modes. Pick one based on the user's needs:
| Mode | When to use |
|---|---|
| plan | Need architecture/design before coding |
| dev | Have an approved plan, ready to implement |
| full | End-to-end with user approval gate between plan and dev |
| auto | End-to-end without approval gate |
| bugfix | Well-specified, reproducible bug (root cause may be unknown) — TDD fix cycle |
Before executing any mode, read the relevant workflow file:
Plan Phase (plan/full/auto) Dev Phase (dev/full/auto) Bugfix Phase (bugfix)
Lead (You) Lead (You) Lead (You)
├── architect [teammate] ├── developer [teammate] ├── tester [teammate]
├── product-manager [teammate] ├── code-tester [teammate] ├── developer [teammate]
└── researcher [subagent] ├── qa-tester [teammate] ├── reviewer [teammate]
├── reviewer [teammate] └── researcher [subagent]
└── researcher [subagent]
│ │
└──── plan.md (handoff) ─────────┘
Teammates message each other directly (Architect teammate↔PM teammate, Developer teammate↔Code-tester teammate, Developer teammate↔QA-tester teammate, Developer teammate↔Reviewer teammate). In bugfix mode: Tester teammate↔Developer teammate, Reviewer teammate↔Developer teammate. Researcher is a subagent — Lead relays results.
Directives mechanism (plan phase): Lead → teammate control signals (e.g. auto_task_baseline, council_review) live in a per-run sidecar JSON file at .dev/cdt/plans/plan-$TIMESTAMP.directives.json, not in prompt prose. See references/directives-schema.md for schema and lifecycle.
Research specialist for doc lookups. Queries Context7 for library docs, searches web for best practices, returns structured findings with code examples. Bundled as agents/researcher.md in this plugin — Context7 MCP is auto-configured via .mcp.json.
Writes the failing regression test BEFORE the developer touches the code (TDD red phase). Verifies the fix passes and re-verifies after refactoring. Iterates directly with developer on failures (max 3 cycles for fix, max 2 for refactor).
Discovers codebase structure using the Explore agent (preferred) or repomix-explorer (if available) for broad understanding, then targets specific files with Glob/Grep/Read for detailed inspection. Reads existing Architecture Decision Records (ADRs) from docs/adrs/ before designing. Designs architecture: components, interfaces, file changes, data flow, testing strategy. Writes new ADRs to docs/adrs/adr-NNNN-<slug>.md for each significant decision. References existing ADRs when relevant and supersedes old ones when decisions change. Debates tradeoffs with PM teammate. Messages design to lead and PM teammate. Writes the plan file as their final deliverable.
Validates architecture against requirements. Challenges design with concerns. Produces verdict: APPROVED or NEEDS_REVISION with specifics.
Implements tasks from plan (dev) or minimal fix from bug spec (bugfix). No stubs, no TODOs. Matches existing patterns. In dev mode: iterates with code-tester teammate on failures, qa-tester teammate on QA issues, reviewer teammate on code quality. In bugfix mode: iterates with tester teammate on failures and reviewer teammate on code quality. Updates project documentation (README.md, AGENTS.md, CLAUDE.md) to reflect implementation changes.
Unit/integration tests. Messages developer teammate with failures + root cause. Max 3 cycles.
Always spawned. Adapts testing approach based on task type: for UI tasks, writes Storybook stories and tests user flows via npx agent-browser; for non-UI tasks, runs integration/smoke tests, verifies regression safety, and tests API contracts. Messages developer teammate with issues + evidence. Max 3 cycles.
Reviews changed files for completeness, correctness, security, quality, and plan adherence (dev) or bug spec adherence (bugfix). Validates review with /council (quick quality for routine, review security or review architecture for critical concerns). Scans for stubs. Messages developer teammate with file:line + fix suggestions. Max 3 cycles. Messages lead with verdict, review cycles, issues found/fixed, and known limitations after approval.
type (impl|test|docs) and depends_on; parallel within waves, sequential betweenYou are a coordinator, not an implementer. During active team phases:
*.ts, *.js, *.py, *.go, *.rs, *.tsx, *.jsx, *.vue, *.svelte, *.css, *.scss, *.html)*.test.*, *.spec.*, __tests__/*) — delegate to code-tester teammate (or tester teammate in bugfix mode)*.md) — delegate to the teammate with context (architect for plans/ADRs, reviewer for reports, developer for project docs)Before replying "standing by" or going idle during an active dev-team session, run TaskList. If any task matches status=pending && blockedBy=[] && owner=null, that task is a wave-gate handoff you owe. Send the kickoff message to the appropriate teammate per dev-workflow.md (test tasks → code-tester / qa-tester per Step 6a; Review all → reviewer per Step 6b) and assign ownership BEFORE going idle.
A pending+unblocked+unowned task during an active dev-team is never "standing by" time — it is always lead action time.
(Plan-team and bugfix-team have different topologies and don't use this same wave-gate handoff shape — their handoffs are covered by their own workflow files.)
npx claudepluginhub p/rube-de-cdt-plugins-cdtReturns a full specification for running a code-mode team with defined roles, rules, research phase, and review protocols.
Sets up multi-agent teams for complex projects with file-based planning, per-agent directories, and teammate spawning. Triggers on team/swarm/start-project requests.
Enforces a strict TDD pipeline for AI agent teams, coordinating feature development through spec, test, implementation, review, and security gates. Activates when spawning parallel agent teams for project work.