deckrd-coder
BDD-style task implementation agent for Deckrd sessions. Implements one task at a time using Red-Green-Refactor cycle. Use when user says "implement task", "code this task", "run deckrd-coder", or provides a Task ID (e.g. T01-02) after /deckrd tasks has been completed. Also use when user says "implement the checklist" or "start BDD implementation". Do NOT use before /deckrd tasks is complete — tasks.md and implementation-checklist.md must exist. Do NOT commit or push — implementation only, no git operations. Do NOT implement multiple tasks in one invocation — one task per call.
From deckrd-codernpx claudepluginhub aglabo/deckrd --plugin deckrd-coderThis skill uses the workspace's default tool permissions.
README.mdassets/languages/go.mdassets/languages/rust.mdassets/languages/shell.mdassets/languages/typescript.mdassets/pressure-scenarios.mdassets/testing-anti-patterns.mdreferences/faq.mdreferences/implementation.mdreferences/troubleshooting.mdreferences/workflow.mddeckrd-coder
Implements one Deckrd task per call using strict BDD: Red → Green → Refactor.
Skill Announcement (REQUIRED)
Before every phase, YOU MUST output:
"I am executing /deckrd-coder [TASK_ID] — Phase [N]: [Phase Name]."
No announcement = violation. Restart with announcement.
Before You Begin (REQUIRED)
Raise ALL questions before writing any code. Ask NOW if any of the following are unclear:
- Task scope or acceptance criteria
- Ambiguous specs in tasks.md
- Implementation approach or dependencies
Once Phase 3 (BDD Implementation) starts, stop asking scope questions. Questions after Phase 3 = Before You Begin was skipped = restart from top.
Usage
/deckrd-coder T01-02 # default checklist
/deckrd-coder T01-02 --checklist <path> # explicit checklist path
Default checklist: tasks/implementation-checklist.md
Execution Flow
deckrd-coder is an orchestration layer. BDD implementation (Red → Green → Refactor) is delegated to the bdd-coder agent per task.
| Phase | Name | What happens |
|---|---|---|
| 0 | Environment | Detect language, test framework, lint, type-check setup |
| 1 | Task Info | Read session, tasks.md, checklist for target task |
| 2 | Dependency Map | Classify tasks into serial / parallel execution groups |
| 3 | bdd-coder Dispatch | Spawn bdd-coder per task; collect status reports |
| 4 | Quality Gate | Global lint + type-check + all tests pass |
| 5 | Done Check | Confirm all checklist items complete |
| 6 | Session End | Reset state; remind user to commit manually |
Gate Rule: phases must run in order. No skipping.
All tests MUST pass at Phase 4. No exceptions. Do NOT commit after completion — user commits manually.
bdd-coder Dispatch (Phase 3)
Pass the following context to each bdd-coder instance:
| Item | Content |
|---|---|
| Task ID | e.g. T-01-02-01 |
| Task description | Full Given/When/Then from tasks.md |
| Quality gate cmds | Commands table from ENV PROFILE |
| Checklist path | Path to implementation-checklist.md |
Do NOT pass: session-wide context, other tasks' info, or session.json.
If bdd-coder reports BLOCKED: stop, report to user, wait for instruction.
References
- Full phase details: workflow.md
- Error recovery: troubleshooting.md
- Q&A: faq.md
- BDD sub-agent: agents/bdd-coder.md
Examples
Implement a single task:
"Implement T01-02." →
/deckrd-coder T01-02
Explicit checklist:
"Use the sprint checklist for T02-01." →
/deckrd-coder T02-01 --checklist docs/.deckrd/my-project/sprint/tasks/implementation-checklist.md
Troubleshooting
tasks.md or checklist not found
Cause: /deckrd tasks has not been run yet.
Solution: Complete the full deckrd flow first: req → spec → impl → tasks.
Tests failing at Phase 4 Cause: bdd-coder implementation is incomplete or incorrect. Solution: Return to Phase 3, re-dispatch bdd-coder for the failing task. Do not skip Phase 4.
Phase skipped accidentally Cause: Announcement not made before a phase. Solution: Restart from the beginning with proper announcements.