From agent-teams-coder
The 7-phase standard development workflow from requirements to delivery. Use when coordinating multi-agent software development tasks.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agent-teams-coder:task-workflowThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
| Phase | Name | Lead Agent | Output |
| Phase | Name | Lead Agent | Output |
|---|---|---|---|
| 1 | Requirements | Marshall | Subtask matrix |
| 2 | Algorithm | Euler | Algorithm design + pseudocode |
| 3 | Development | Forge | Source code |
| 4 | Testing | Sentinel | Test report |
| 5 | Analysis | Lens | Code analysis report |
| 6 | Documentation | Atlas | Software manual (4 chapters) |
| 7 | Delivery | Marshall | Consolidated deliverables |
Phase 1 ──→ Phase 2 ──→ Phase 3 ──→ Phase 4 ──→ Phase 5 ──┐
↑ │ │
└─ Bug ───┘ ▼
Phase 6 ──→ Phase 7
| Category | Phases Used | Skip |
|---|---|---|
| New feature | All 7 | None |
| Bug fix | 1, 3, 4, 7 | 2 (algorithm), 5 (analysis), 6 (docs) |
| Documentation update | 1, 5, 6, 7 | 2, 3, 4 |
| Performance optimization | 1, 2, 3, 4, 7 | 5 (analysis), 6 (docs) |
Every agent MUST notify Chronicle upon completing their phase work:
bash scripts/notify.sh <agent_name> chronicle "<phase summary>" "<details: what was done, files produced, key data>"
| Phase | Who CCs Chronicle | Content |
|---|---|---|
| 2 | Euler | Algorithm design summary, complexity analysis |
| 3 | Forge | Code files, language, lines, implementation notes |
| 4 | Sentinel | Test case count, pass rate, bug list, fix status |
| 5 | Lens | Analysis scope, issues found, architecture assessment |
| 6 | Atlas | Document chapters, page count, coverage |
| 7 | Marshall | Final deliverables list, overall summary |
Chronicle cannot passively "listen" — it must be actively notified by each member.
npx claudepluginhub xinjing-guo/agent-teams-coder --plugin agent-teams-coderProvides structured workflows for building, modifying, or fixing software/tools/systems with clear requirements, using composable phases and paradigms for greenfield, enhancement, bugfix, and addition tasks.
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.