From story-coding-agent
Runs after each domain agent completes. Extracts a structured summary from the agent's output by scanning files modified during its run. The summary is injected into the next domain agent's context packet so downstream agents have concrete artifact references (table names, API endpoint URLs, component names) rather than relying on unstructured text.
How this skill is triggered — by the user, by Claude, or both
Slash command
/story-coding-agent:context-bridgeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Extracts structured handoff artifacts from a completed domain agent's run and formats them for injection into the next domain agent's context. This ensures cross-domain implementation consistency — when the backend agent runs after the DB agent, it knows exactly which tables, columns, and migration files were created.
Extracts structured handoff artifacts from a completed domain agent's run and formats them for injection into the next domain agent's context. This ensures cross-domain implementation consistency — when the backend agent runs after the DB agent, it knows exactly which tables, columns, and migration files were created.
| Input | Source | Description |
|---|---|---|
completedAgent | domain-invoker | Name of the domain agent that just finished |
agentOutput | domain-invoker | Raw output and file list from the domain agent's run |
ticket | ticket-fetch | Story ticket for context |
Use Glob and Read tools to scan files written or modified during the domain agent's run. Extract domain-specific signals by file type and content pattern.
db-coding-agent)Scan: migration files (migrations/*.sql, **/migrations/**), ORM model files, schema files.
Extract:
CREATE TABLE, ALTER TABLE statements or model class namesCREATE INDEX statementsREFERENCES clauses or FK decoratorsuser_sessions with 4 columns. Migration: migrations/20260324_add_user_sessions.sql."backend-coding-agent)Scan: route files, controller files, service files, DTO/schema files.
Extract:
@Get('/path'), router.get('/path'), app.get('/path'))process.env.* or config referencesfrontend-coding-agent)Scan: component files, page files, route config files, state management files.
Extract:
export function ComponentName or export default@Vensure-Devops-QA/prismatic-*devops-coding-agent)Scan: pipeline YAML files, Dockerfile, .env.example files.
Extract:
.env.examplesecurity-coding-agent)Scan: middleware files, auth files, validation files, audit log files.
Extract:
If file scanning cannot produce structured artifacts (e.g., agent wrote no files, or output is purely in-memory), use the agent's text output as the summary directly, prefixed with "(unstructured)".
{
"agent": "db-coding-agent",
"summary": "Created table `user_sessions` with columns: id (UUID PK), user_id (FK → users.id), token (VARCHAR 512), expires_at (TIMESTAMP). Migration: migrations/20260324_add_user_sessions.sql.",
"artifacts": [
"migrations/20260324_add_user_sessions.sql",
"src/models/UserSession.ts"
],
"contextForNext": "The next agent should reference `user_sessions` table for JWT session storage. FK constraint on `users.id` is enforced at the DB layer. Column names: id, user_id, token, expires_at."
}
| Field | Description |
|---|---|
agent | Name of the domain agent that produced these artifacts |
summary | One or two sentence human-readable summary of what was implemented |
artifacts | Relative paths of files written or materially modified during this agent's run |
contextForNext | Explicit instructions for the next agent referencing concrete names (table names, endpoint URLs, component names) produced by this agent |
The contextForNext field is what gets injected into the next agent's completedAgents[] entry. It must contain concrete names — not vague descriptions — so the downstream agent can reference them directly in its implementation.
npx claudepluginhub gagandeepp/software-agent-teams --plugin story-coding-agentGuides completion of development work by verifying tests, detecting environment, and presenting structured options for merge, PR, or cleanup.
Enforces test-driven development: write failing test first, then minimal code to pass. Use when implementing features or bugfixes.
Guides creation and editing of skills using test-driven development with pressure scenarios and subagents to verify agent compliance.