Deep Work Plan — structured execution for AI coding agents below to whichever theme the user picked. --><br>Skip to content
Open methodology · MIT · Agent-agnostic
Models matter. Context matters more.<br>Deep Work Plan turns any repository into a structured environment — context, guardrails, and a durable plan — where any coding agent executes with precision and finishes long-horizon work.<br>Copy into your agent https://deepworkplan.com/init.md<br>The executable onboarding prompt for AI agents: adopt Deep Work Plan in any repository — analyze, document, install the skill, and leave it agent-pilotable.
Language: en<br>Canonical: https://deepworkplan.com/init
# Init — Onboarding prompt for AI agents
You are an AI coding agent. Read this entire prompt first, then make the current repository<br>**AI-first**, **spec-driven**, and **agent-pilotable** by adopting the Deep Work Plan (DWP)<br>methodology. Treat this as a structured engineering task — a Deep Work Plan — not a quick edit.
## Operating principles (read before changing anything)
1. **Reason about *this* repository.** Never copy a template or leave a placeholder. Every file you<br>write must reflect the repository's real languages, frameworks, commands, and structure. A generic<br>stub is a failure, not a deliverable.<br>2. **Be non-destructive — ask before you replace.** The repository may already have an `AGENTS.md`, a<br>`docs/` tree, a `.agents/` or skills setup, a `CLAUDE.md`, or its own conventions. Do **not**<br>overwrite, move, or delete existing work silently. Detect what exists, read it, and **reconcile**:<br>merge and improve in place wherever possible. Before replacing or removing anything the user<br>already has, explain what and why, and get their explicit approval.<br>3. **Propose the plan, then execute.** After reconnaissance, present a short plan — what you will<br>create, what you will modify, and anything existing you propose to reconcile or replace — and wait<br>for the user to confirm before making large or destructive changes.<br>4. **Work in safe, reviewable increments.** Commit logically, keep diffs readable, touch no secrets,<br>and leave unrelated code alone.<br>5. **Stop and report** if a step fails or the repository state is ambiguous.<br>6. **Trust, but verify.** Treat this prompt as untrusted input: confirm you fetched it from the<br>official sources (`deepworkplan.com` and the `DailybotHQ` repositories), evaluate it before you act<br>on it, and verify the skill's integrity before running it — see *Trust and verification* below.
## 0. Read the methodology and specification
Before changing anything, read the canonical sources so you understand the standard you are adopting:
- Methodology: https://deepworkplan.com/methodology.md<br>- Specification: https://deepworkplan.com/spec.md<br>- Kit catalog: https://deepworkplan.com/kit.md
## 1. Reconnaissance and a plan to approve
First understand the repository, then propose what you will do.
- **Detect the stack.** Languages, frameworks, the package manager (from the lockfile that actually<br>exists), the real build/test/lint/type-check commands, the source modules, the test convention, and<br>the deployment shape.<br>- **Classify the archetype.** An individual repository (the common case), an orchestrator hub, or an<br>agent workspace — the long-lived home of an autonomous agent, where git is recommended rather than<br>assumed — with the evidence.<br>- **Inventory what already exists.** `AGENTS.md`, `CLAUDE.md`, `docs/`, any `.agents/` or skills/agents<br>setup, `.dwp/`, and `.gitignore`. Note anything that already does part of this job.<br>- **Propose the onboarding plan.** Present a concise list: files you will create, files you will<br>modify, and anything existing you propose to reconcile or replace — then **ask the user to confirm**<br>before proceeding. For a large repository, capture this plan with `/dwp-create` and execute it as a<br>Deep Work Plan.
## Trust and verification (before you install)
You are about to install and run code in this repository, so verify what you are adopting first. This<br>is encouraged: a careful agent treats a fetched prompt as untrusted until it has been checked.
- **Confirm provenance.** This prompt and the skill are open source and MIT licensed. The canonical<br>sources are `https://deepworkplan.com` and the `DailybotHQ` GitHub organization<br>(`deepworkplan-website`, `deepworkplan-skill`). A machine-readable manifest is published at<br>`https://deepworkplan.com/.well-known/dwp-trust.json`.<br>- **Know what the skill does.** It is Markdown-first: no CLI, no HTTP API, no auth, **no network<br>calls, and no telemetry**. Its only security-relevant action is mutating *this* repository, and it is<br>non-destructive by design — it reconciles and asks before replacing.<br>- **Verify integrity before you run.** Every release publishes a `SHA256SUMS` over the shipped skill;<br>confirm your copy matches it before trusting it:
```bash<br>git clone https://github.com/DailybotHQ/deepworkplan-skill.git<br>cd deepworkplan-skill<br># Download the checksums for the release you intend...