Claude Skill MartBrowse skillsWhat is a Skill?
 
← Back to list

ULW Loop (Evidence-Bound Goal Execution)

A durable goal-execution loop that decomposes long work into atomic steps and closes them only with real, observable evidence.

AutomationAdvanced★ 67,531⑂ 5,507AI score 6/10Last updated: Aug 9, 2026

What it does

ulw-loop runs long-horizon engineering work as a strict loop: goals → ultra-granular steps → recorded evidence.

  • Registers goals up front in the omo ulw-loop CLI state (.omo/ulw-loop) and mirrors every action into a live checklist with exactly one step in progress.
  • Refuses "tests passed" as proof of done. Each success criterion needs evidence observed on a real surface — terminal/TUI, HTTP, browser, computer-use, or auxiliary surfaces like CLI stdout, DB diffs, parsed config dumps.
  • Binds evidence to the git tree hash, so it only goes stale when tracked content actually changes; otherwise you re-run at current HEAD and re-record rather than relabel.
  • Survives compaction: re-read brief, goals, ledger and omo ulw-loop status --json before resuming, never re-plan from scratch.
  • Delegates code edits, tests and QA to right-sized Codex subagents, with a strict TASK/DELIVERABLE/SCOPE/VERIFY message contract, background spawning, and backoff on waits.
  • Treats team mode as a deliberate decision: only when work units overlap unavoidably and parallelism truly finishes sooner, with one git worktree per member and per-unit merges.

Who it's for

  • Developers running multi-agent, long-running tasks inside Codex / oh-my-openagent (omo).
  • Teams that keep getting bitten by "green CI, broken product".
  • Anyone managing large refactors or migrations that outlive a single context window.

Examples

  1. Large migration: "Use ulw-loop to move the billing module to the v2 API" — goals registered, atomic commits per verified unit, HTTP/terminal evidence attached.
  2. Feature with manual QA: implement a screen, then verify it in the browser channel before any step is marked complete.
  3. Parallel coordination: overlapping units get per-member worktrees and branches, each landing as soon as its own gates go green, with the lead sequencing conflicts.

· · · Install guide · · ·

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 Claude Skill Mart.
Copy the packages/omo-codex/plugin/components/ulw-loop/skills/ulw-loop folder from the GitHub repo code-yeongyu/oh-my-openagent into my ~/.claude/skills/ulw-loop/.
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/omo-codex/plugin/components/ulw-loop/skills/ulw-loop ~/.claude/skills/

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

  1. Open a terminal.
  2. Clone the repository: git clone https://github.com/code-yeongyu/oh-my-openagent.git
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy the skill: cp -r oh-my-openagent/packages/omo-codex/plugin/components/ulw-loop/skills/ulw-loop ~/.claude/skills/
  5. Confirm references/full-workflow.md came along — the actual workflow lives there.
  6. Install and configure oh-my-openagent (omo), since the skill drives the omo ulw-loop CLI and Codex multi-agent tools.
  7. Restart Claude Code and try a prompt like "run this with ulw-loop" to confirm the skill triggers.
View source on GitHub ↗License: NOASSERTION