skills/task-plan/SKILL.md
Defines the canonical plan.md format used as a handoff document between agents and across context boundaries. Use this skill whenever creating a new task plan, resuming work on a task with an existing plan, updating an in-progress plan, or when writing a plan that another agent (or yourself after context reset) needs to pick up cold. If a task has 3+ steps or spans multiple sessions, you need a plan.md and this skill tells you the format.
npx skillsauth add maestria-co/ai-playbook task-planInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
plan.md is a handoff document — it contains enough context for any agent
(or the same agent after a context clear) to resume work without re-reading the
entire codebase. It is not a progress tracker or a to-do list.
This skill defines the canonical format and rules for maintaining plan files.
Every plan.md lives in .context/tasks/[TASK-ID]/plan.md and follows this structure:
# [TASK-ID]: [Short Title]
## Branch
`feature/TASK-ID-short-description`
## Objective
One paragraph. What are we building and why?
## Acceptance Criteria
- [ ] Criterion 1
- [ ] Criterion 2
- [ ] Criterion 3
## Decisions
Key decisions made during planning or implementation.
- **Decision:** [What was decided]
**Rationale:** [Why]
## Key Files
Files central to this task — helps the next agent load context fast.
- `path/to/file.ts` — [role in this task]
- `path/to/other.ts` — [role in this task]
## Task Breakdown
Ordered steps to complete the work.
1. [x] Step 1 — completed description
2. [ ] Step 2 — in progress description ← CURRENT
3. [ ] Step 3 — upcoming description
## Progress Log
Reverse-chronological entries. One per work session.
- **[Date/Session]:** What was accomplished, what's next
## Blockers
- [Blocker description] — [status: investigating / escalated / resolved]
- [ ] / - [x])[x] only when verified with evidence.context/decisions/ for major decisions← CURRENT[x]| Situation | Create plan? |
| --------------------------------- | ------------------------------------- |
| Multi-step feature (3+ steps) | Yes |
| Bug fix requiring investigation | Yes — helps track debugging progress |
| Simple one-file change | No — overkill |
| Refactoring across multiple files | Yes |
| Research task | No — use .context/research/ instead |
Update the plan whenever:
Before continuing work on any step, verify the plan reflects current reality:
If the plan is stale (e.g., steps were skipped, scope changed, a decision was made without recording it):
[x], add decisions, update ← CURRENTRule: A plan that doesn't match current state is worse than no plan — it actively misleads the next agent. Fix it before using it.
When resuming a task from a plan:
← CURRENT marker.context/tasks/[TASK-ID]/plan.mddevelopment
Writes and runs a test suite for a piece of code, covering happy path, edge cases, error cases, and security cases. Use when: implementation is complete and needs test coverage, a bug needs a reproduction test and fix validation, or code needs coverage before a refactor. Do not use when: the code under test is not yet implemented, or the spec is still unclear.
testing
Use when creating a new skill, editing an existing skill, or helping a user author a skill for this system. Covers structure, discoverability, quality, and discipline hardening.
development
Evidence-based verification process to run before marking any task complete. Use this skill every time you're about to report that work is done — for features, bug fixes, refactoring, or any code change. This catches the most common failure mode: declaring "done" without proof. If you're finishing up and about to tell the user the task is complete, run this checklist first.
development
Teaches agents how to discover, select, and invoke skills from the skill library. Use this skill whenever you're uncertain which skill applies to a task, when composing multiple skills for complex work, or when you need to understand what skills are available. This is your go-to when facing an ambiguous task and need to figure out the right approach before diving into implementation.