para-help

name	para-help
description	Display the PARA-Programming quick-reference guide with all 11 skills and the Research→Plan→Review→Execute→Review→Summarize→Archive workflow.
model	haiku
effort	low

Display the PARA-Programming quick-reference guide.

Usage

para-help

Implementation

Determine how to render user-facing skill requests before displaying the reference:

1. Render a client selector only when the host is explicit in runtime/system context or the user names the client.
2. Do not infer the host from ~/.agents/skills or another shared discovery path.
3. For an unknown client, OpenCode, or Gemini CLI, use exact-name natural language such as: Use the para-plan skill to plan <task>.
4. Treat the forms below as presentation only. The canonical request remains para-<skill> [arguments].
5. Preserve the exact para-* identifier on every line. Do not replace identifiers with generic labels such as “the plan review skill” or reorder/expand the Typical Flow.

Client	User-facing form
OpenAI Codex	$para-<skill> [arguments]
Cursor	/para-<skill> [arguments]
Pi, skill commands enabled	/skill:para-<skill> [arguments]
Pi, skill commands disabled	natural-language — Use the para-<skill> skill <intent-and-arguments>.
OpenCode	natural-language — Use the para-<skill> skill <intent-and-arguments>.
Gemini CLI	natural-language — Use the para-<skill> skill <intent-and-arguments>.
Unknown client	natural-language — Use the para-<skill> skill <intent-and-arguments>.

Display the reference below. Keep the skill identifiers canonical. In the Typical Flow, render each request using the selected client form; when using natural language, include both the exact skill name and its intent/arguments.

For natural-language clients, render the Typical Flow with this exact request sequence:

Use the `para-research` skill to research "Add user authentication".
Use the `para-plan` skill to plan "Add user authentication".
Use the `para-review` skill with `--plan`.
Use the `para-workflow` skill. Or manually:
  Use the `para-execute` skill.
  Use the `para-review` skill with `--pr`.
  Use the `para-summarize` skill.
  Use the `para-archive` skill.

---

PARA-Programming Quick Reference

Workflow: Research → Plan → Review → Execute → Review → Summarize → Archive

Detailed workflow: Research → Plan → Review Plan → Execute → Review PR → Summarize → Archive

When to Use PARA

Use PARA if the task results in git changes: features, bug fixes, refactoring, config, migrations, tests, documentation edits, or complex debugging.

Skip PARA if the task is read-only or informational: questions, navigation, explanations, or state inspection.

Skills

Skill	Purpose
para-init	Initialize PARA structure in a project
para-research <task>	Deep codebase research before planning
para-plan <task>	Create a planning document through collaboration
para-review --plan|--pr	Staff+ review loop for plans and PRs
para-execute	Create worktree, extract todos, start execution
para-workflow	Orchestrate execute → PR → review → summarize → archive
para-summarize	Generate post-work summary
para-archive	Archive context and start fresh
para-status	Check current workflow state
para-check	Decide whether a request needs PARA
para-help	Show this reference

Typical Flow

para-research Add user authentication
para-plan Add user authentication
para-review --plan
para-workflow           # or manually:
  para-execute          #   -> Creates worktree, implements with TDD
  para-review --pr      #   -> Staff+ review loop
  para-summarize        #   -> Generate summary
  para-archive          #   -> Clean up worktree

File Structure

context/
├── context.md       # Active session context
├── plans/           # YYYY-MM-DD-task-name.md
├── summaries/       # YYYY-MM-DD-task-name-summary.md
├── archives/        # YYYY-MM-DD-context.md
├── data/            # Input/output files, research docs
└── servers/         # MCP tool wrappers

Tips

• Use the para-status skill to see where you are in the workflow
• Use the para-check skill if unsure whether a task needs PARA
• Full methodology details are in ../../docs/METHODOLOGY.md