CommunityProductivity & Collaborationgithub.com

aitormf/agent-skills

Splits an approved technical plan into small tasks, ordered by dependencies, with affected files, per-task verification and parallel groups. Use it when there is a plan and someone asks to break it down into tasks or get it ready for implementation.

What is agent-skills?

agent-skills is a Claude Code agent skill that splits an approved technical plan into small tasks, ordered by dependencies, with affected files, per-task verification and parallel groups. Use it when there is a plan and someone asks to break it down into tasks or get it ready for implementation.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/aitormf/agent-skills/tree/HEAD/skills/plan-to-tasks

Installed? Explore more Productivity & Collaboration skills: steipete/gemini, steipete/gh-issues, steipete/skill-creator · View all 6 →

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

From plan to tasks

Turn a plan into an executable task list that the implement agent can follow and tick off.

  • Each work item lives in docs/specs/<YYYYMMDD>-<type>-<name>/.
  • tasks.md has no status: its progress is the [x] checkboxes.

Before you start

  1. Identify the work item and read plan.md and spec.md.

  2. Gate: if the plan does not have status: approved, stop and ask:

    The plan is <status>. Shall I continue anyway, review it first with review-spec, or will you approve it now?

    Continue only with an explicit answer. For a small change with no plan, the gate applies to the spec.

  3. Read the verification command in the Commands section of AGENTS.md (Comandos in Spanish-language projects). If there is none, ask what it is or leave the wrap-up task with a warning.

How to split

  • Small tasks: each one is a reviewable change that leaves the project green.
  • Ordered by dependencies. Mark with [P] those that can be done in parallel with the previous one.
  • Each task lists the files it touches, if known, and how it is verified: which test, or which command.
  • If the project works with TDD (CONTRIBUTING.md or AGENTS.md will say so), the test goes before the code or in the same task, never at the end.
  • Every acceptance criterion of the spec and every negative test from Security must be covered by at least one task. If one is not, say so.
  • The Wrap-up phase is mandatory: acceptance criteria with tests and the verification command green.
  • Do not add work that is not in the plan. If something is missing from the plan, raise it as a question.

Result

  1. Draft tasks.md from assets/tasks-template.md.
  2. Show the draft with:
    • the tasks;
    • the parallel groups;
    • the verification points;
    • any remaining questions.
  3. Write it only once the person approves.
  4. Suggested next step: implement it with the implement agent.

Individual skills in this repo

This repo contains 3 individual skills — each has its own dedicated page.

Related Skills