Menú
Use when generating or validating the ExecutionPlan JSON that the orchestrator must produce before spawning any agents
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Basado en la clasificación ocupacional SOC
| name | execution-plan |
| description | Use when generating or validating the ExecutionPlan JSON that the orchestrator must produce before spawning any agents |
The orchestrator MUST always produce a valid ExecutionPlan JSON before spawning any agents. This is the contract between the orchestrator and the rest of the system.
{
"run_id": "string",
"task": "string",
"target_dir": "string | null",
"task_brief": {
"type": "string (frontend|backend|fullstack|infra|fix|other)",
"complexity": "string (low|medium|high)",
"inferred": {},
"explicit_requirements": ["string"],
"implicit_requirements": ["string"],
"ambiguities": ["string"],
"out_of_scope": ["string"]
},
"agents": [
{
"role": "string",
"model": "string (haiku|sonnet|opus)",
"skills": ["string"],
"mcps": ["string"],
"depends_on": "string | string[] | null",
"worktree": "string | null",
"context": "string"
}
]
}
| Field | Required | Description |
|---|---|---|
run_id | yes | Unique identifier: YYYY-MM-DD-NNN format |
task | yes | Original user input, verbatim |
target_dir | yes | Where generated code goes. projects/{name} in monorepo mode, . in external mode |
task_brief | yes | Structured breakdown produced by brainstorm agent (or orchestrator directly for low complexity) |
task_brief.type | yes | Task domain classification |
task_brief.complexity | yes | Used to decide model selection and whether brainstorm was needed |
agents[] | yes | Ordered list of agents to spawn. At least 1 required. |
agents[].role | yes | Descriptive name: brainstorm, backend-developer, frontend-developer, db-architect, pr-creator, pr-reviewer, synthesizer |
agents[].model | yes | Model to use. See catalog/models.md for selection rules |
agents[].skills | yes | Skills to inject. Empty array [] if none needed |
agents[].mcps | yes | MCPs available to this agent. Empty array [] if none needed |
agents[].depends_on | yes | null = runs immediately. String = waits for that role. Array = waits for all |
agents[].worktree | conditional | Required when agent writes code. Branch name for isolated execution. null for read-only agents |
agents[].context | yes | What this agent needs to know. Include outputs from dependency agents |
depends_on: null run immediately and in parallel with each otherdepends_on: "role-name" wait for that role to completedepends_on: ["role-a", "role-b"] wait for ALL listed roles (barrier)synthesizer always runs last — set depends_on to all other agent roles{
"run_id": "2026-04-06-001",
"task": "add JWT auth to the FastAPI backend",
"target_dir": "projects/todo-app",
"task_brief": {
"type": "backend",
"complexity": "medium",
"inferred": {
"auth_type": "JWT",
"reason": "FastAPI stack detected, no OAuth mentioned"
},
"explicit_requirements": ["login endpoint", "refresh token"],
"implicit_requirements": ["users table", "password hashing", "rate limiting"],
"ambiguities": [],
"out_of_scope": ["OAuth providers", "2FA"]
},
"agents": [
{
"role": "backend-developer",
"model": "sonnet",
"skills": ["fastapi-patterns", "security-patterns"],
"mcps": ["filesystem"],
"depends_on": null,
"worktree": "feat/jwt-auth",
"context": "Implement JWT auth per the task_brief. Use FastAPI dependency injection for auth middleware."
},
{
"role": "pr-creator",
"model": "haiku",
"skills": [],
"mcps": ["filesystem", "github"],
"depends_on": "backend-developer",
"worktree": null,
"context": "Create PR from feat/jwt-auth branch. Use outputs from backend-developer in context.json."
},
{
"role": "pr-reviewer",
"model": "sonnet",
"skills": ["security-patterns"],
"mcps": ["github"],
"depends_on": "pr-creator",
"worktree": null,
"context": "Review the PR for security issues. Focus on JWT implementation, token storage, and rate limiting."
},
{
"role": "synthesizer",
"model": "sonnet",
"skills": [],
"mcps": [],
"depends_on": ["pr-reviewer"],
"worktree": null,
"context": "Aggregate all outputs from context.json and produce final run summary."
}
]
}
Before finalizing the ExecutionPlan, verify:
roledepends_on references valid role names in the same plansynthesizer is always lastworktree setcontext field is specific — not generic fillerOperate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing. Use when AI agents perform most implementation work and humans enforce quality and risk controls.
REST API design patterns including resource naming, status codes, pagination, filtering, error responses, versioning, and rate limiting for production APIs.
Deployment workflows, CI/CD pipeline patterns, Docker containerization, health checks, rollback strategies, and production readiness checklists for web applications. Use when setting up deployment infrastructure or planning releases.
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML/CSS layouts, or when styling/beautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.
Research-before-coding workflow. Search for existing tools, libraries, and patterns before writing custom code. Systematizes the "search for existing solutions before implementing" approach. Use when starting new features or adding functionality.
Use when an agent writes, reviews, or audits code that handles authentication, authorization, user input, or sensitive data