OpenParable
Skip to content
← Back to skills

Make Plans

Productivity
make-plans

Make Plans is an Agent Skill for productivity published in the GetStream/stream-chat-react repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Structures plan.md files for parallel, non-overlapping agent work. Use when creating or editing execution plans, task lists for multiple agents, or when the user asks how to structure plans or split work into tasks.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/GetStream/stream-chat-react/blob/HEAD/.cursor/skills/make-plans/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/make-plans, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Making plan.md Files

Plans enable parallel, non-overlapping work. Each task must be independent; agents work in a dedicated git worktree (see worktrees skill).

Where to keep plan.md and spec files: Create and edit plan.md (and the other Ralph spec files: goal.md/spec.md, state.json, decisions.md) in a dedicated folder under the specs folder at the repo root: specs/<feature-or-component-name>/. Example: specs/resizable-container/plan.md. Do not place spec files next to source (e.g. not under src/components/...).

Required plan.md Structure

  1. Worktree section (top) — path, branch, base branch. See worktrees skill.
  2. Task overview — one line: tasks are self-contained; same-file tasks have dependencies.
  3. Task sections — each with the template below.
  4. Execution order — phases showing what can run in parallel.
  5. File ownership summary — table: Task | Creates/Modifies.

Task Template

Every task must include:

## Task N: <Descriptive Name>

**File(s) to create/modify:** `path/to/file.tsx`

**Dependencies:** Task X (or "None")

**Status:** pending | in-progress | done | blocked

**Owner:** <agent-name> | unassigned

**Scope:**

- Bullet points; include interfaces/snippets if helpful

**Acceptance Criteria:**

- [ ] Verifiable condition
- [ ] e.g. TypeScript compiles, exports from module

Task Independence Rules

  • Same file → cannot run in parallel. Add explicit dependencies so only one agent touches a file at a time.
  • Explicit dependencies: If B needs A, write **Dependencies:** Task 1.
  • Self-contained scope: Task must be doable without other in-progress tasks (include types, imports, exports).
  • Same-file options: Either one integration task (e.g. "Update index.tsx for all components", deps: 1–4) or a chain (Task 7 adds A, Task 8 adds B with dep Task 7).

Task Granularity

  • Too coarse: "Implement entire Gallery" — not parallelizable.
  • Too fine: Separate tasks for "create file", "add imports", "add type" — overhead.
  • Good: One task = one logical unit (e.g. "GalleryContext: types + context + hook", "Gallery Provider: state + navigation", "GalleryUI: rendering + interactions").

Cross-Cutting / Integration

Use a final integration task, e.g.:

  • Update index.tsx exports for all new components.
  • Update main stylesheet to import component styles.
  • Dependencies: all preceding tasks.

Checklist Before Finalizing plan.md

  • Worktree section at top
  • Same-file tasks have dependencies
  • Dependencies explicit; execution order and file ownership table present
  • Tasks sized appropriately; each has Status, Owner, acceptance criteria

For full examples and worktree setup, see reference.md and the worktrees skill.