AI Skill MartBrowse skillsPromptsProgramsLearnWhat is a Skill?
 
← Back to list

ulw-plan — Explore-First Planning Consultant

A planning-only skill that grounds in your codebase, asks just the decisions you must own, and writes one decision-complete work plan after your explicit approval.

Dev & CodingAdvanced★ 67,531⑂ 5,507AI score 9/10Last updated: Aug 9, 2026

What it does

  • Puts the assistant into a 'Prometheus' planning-consultant role that never implements. "just do it" is interpreted as "plan it", and delegated implementation is blocked too.
  • Routes intent as CLEAR vs UNCLEAR: clear goals get a short interview limited to genuine owner-decisions; fuzzy goals get research plus announced best-practice defaults instead of an interrogation.
  • Runs parallel read-only exploration first (codegraph / Read / Grep / Glob / external docs) so discoverable facts are researched, never asked.
  • Enforces an approval gate: one short brief, then it waits for your explicit okay. Approval authorizes writing the plan only — execution happens in a separate worker session ($start-work).
  • Uses scripts/scaffold-plan.mjs to create .omo/drafts/<slug>.md (a compaction-safe resume point) and, after approval, .omo/plans/<slug>.md, with strict task grammar (- [ ] 1. title, final verifiers - [ ] F1. title) and a structural self-check before handoff.
  • Any "high accuracy / deep review" phrasing flips on a mandatory dual review (momus + Oracle).

Who it's for

  • Developers who want a zero-judgment execution plan before handing a large change to an AI worker.
  • Teams that need a structural guardrail against the model editing code mid-conversation.
  • Users following a plan → approve → execute split (OpenAgent/OpenCode-style tooling).

Example uses

  1. ulw-plan + "add a 5/min-per-IP rate limit to /login" → routed CLEAR, existing middleware explored, only owner-level forks asked (counter store, error shape), plan written after approval.
  2. "make auth better" → routed UNCLEAR, no extra questions, researched defaults announced and the high-accuracy review runs automatically.
  3. "break this epic down, high accuracy" → numbered todos plus a final verification wave, each todo carrying reference files, acceptance criteria, agent-run QA, and a commit boundary.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/code-yeongyu/oh-my-openagent/HEAD/packages/shared-skills/skills/ulw-plan/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on AI Skill Mart.
Copy the packages/shared-skills/skills/ulw-plan folder from the GitHub repo code-yeongyu/oh-my-openagent into my ~/.claude/skills/code-yeongyu-ulw-plan-3/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/code-yeongyu/oh-my-openagent.git && mkdir -p ~/.claude/skills && cp -r oh-my-openagent/packages/shared-skills/skills/ulw-plan ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal in your working directory.
  2. Clone the repo: git clone https://github.com/code-yeongyu/oh-my-openagent.git
  3. Create the skills folder: mkdir -p ~/.claude/skills
  4. Copy the whole skill folder (keeping references/ and scripts/): cp -r oh-my-openagent/packages/shared-skills/skills/ulw-plan ~/.claude/skills/
  5. Verify node --version (or bun --version) — the scaffold script needs one of them.
  6. Restart Claude Code and type ulw-plan or "plan this before coding". You should see ULW-PLAN MODE ENABLED! as the first line.
  7. Find generated artifacts under .omo/drafts/ and .omo/plans/ in your project root; consider adding .omo/ to .gitignore.
View source on GitHub ↗License: NOASSERTION