ulw-loop
Start an ULTRAWORK loop — Ralph loop plus mandatory Oracle verification before exit.
Skill by mihazs
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/mihazs-ulw-loop
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: productivity, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/ulw-loop
What's inside
- Skill name
ulw-loop- Description
- Start an ULTRAWORK loop — Ralph loop plus mandatory Oracle verification before exit. Use for /ulw-loop, /ultrawork, or "ultrawork <task>". Ends only after <promise>VERIFIED</promise> from the verifier subagent. Cancel with /cancel-ralph.
- License
- MIT
- Instructions
- 35 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# ULTRAWORK Loop ## Start ```text /ulw-loop "task description" [--completion-promise=DONE] [--max-iterations=500] /ultrawork "task description" ultrawork refactor the payment module ``` Default max iterations: **500** (vs 100 for `/ralph-loop`). ## Flow 1. Work until fully done → output `<promise>DONE</promise>` (not final). 2. Stop hook enters **verification** — you must run a verifier subagent (default: `code-reviewer`). 3. Verifier must end with: ```text Agent: oracle <promise>VERIFIED</promise> ``` 4. Only then does the loop clear and the session may stop. ## If verification fails Fix issues, emit `<promise>DONE</promise>` again, and re-run verification. ## Cancel `/cancel-ralph` (same as Ralph loop — one state file). ## State `.omg/ralph-loop.local.md` with `ultrawork: true` and `verification_pending` when awaiting verification.
From mihazs/oh-my-grok, MIT.
More from mihazs/oh-my-grok
- 0.542score
ralph-loop
mihazs
Start a Ralph Loop — autonomous work-until-done via Stop-hook continuations.
SkillCFreshness: C★ 11Claude Code +9 - 0.542score
handoff
mihazs
Create a detailed HANDOFF CONTEXT summary for continuing work in a new Grok session.
SkillCFreshness: C★ 11Claude Code +9 - 0.542score
agent-skill-gate
mihazs
MANDATORY before code changes, debugging, planning, or multi-step implementation in ANY repository.
SkillCFreshness: C★ 11Claude Code +9 - 0.539score
lsp
mihazs
Language Server Protocol tools via the bundled lsp MCP server — diagnostics, navigation, symbols, and rename.
SkillCFreshness: C★ 11Claude Code +9 - 0.539score
prometheus-plan
mihazs
Strategic planning mode (/plan): interview, gap analysis, write work plans under .omg/plans/, optional review, then /start-work to activate boulder execution.
SkillCFreshness: C★ 11Claude Code +9 - 0.539score
ast-grep
mihazs
Structural code search and rewrite via the bundled ast_grep MCP server.
SkillCFreshness: C★ 11Claude Code +9 - 0.539score
hashline-edit
mihazs
Hash-anchored StrReplace edits using LINE#ID tags from Read output.
SkillCFreshness: C★ 11Claude Code +9 - 0.537score
cancel-ralph
mihazs
Cancel an active Ralph or Ultrawork loop.
SkillCFreshness: C★ 11Claude Code +9 - 0.517score
finishing-a-development-branch
mihazs
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting…
SkillCFreshness: C★ 11Claude Code +9 - 0.516score
requesting-code-review
mihazs
Use when completing tasks, implementing major features, or before merging to verify work meets requirements
SkillCFreshness: C★ 11Claude Code +9 - 0.516score
subagent-driven-development
mihazs
Use when executing implementation plans with independent tasks in the current session
SkillCFreshness: C★ 11Claude Code +9 - 0.516score
writing-plans
mihazs
Use when you have a spec or requirements for a multi-step task, before touching code
SkillCFreshness: C★ 11Claude Code +9
Signals
- Stars
- 11
- Forks
- 5
- Open issues
- 0
- Last commit (this folder)
- Jun 2, 2026 (4 months ago)
- Last release
- Jun 3, 2026 (4 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/mihazs/oh-my-grok
- Website
- github.com/mihazs/oh-my-grok
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from mihazs
- 0.542score
ralph-loop
mihazs
Start a Ralph Loop — autonomous work-until-done via Stop-hook continuations.
SkillCFreshness: C★ 11Claude Code +9 - 0.542score
handoff
mihazs
Create a detailed HANDOFF CONTEXT summary for continuing work in a new Grok session.
SkillCFreshness: C★ 11Claude Code +9 - 0.542score
agent-skill-gate
mihazs
MANDATORY before code changes, debugging, planning, or multi-step implementation in ANY repository.
SkillCFreshness: C★ 11Claude Code +9 - 0.539score
lsp
mihazs
Language Server Protocol tools via the bundled lsp MCP server — diagnostics, navigation, symbols, and rename.
SkillCFreshness: C★ 11Claude Code +9 - 0.539score
prometheus-plan
mihazs
Strategic planning mode (/plan): interview, gap analysis, write work plans under .omg/plans/, optional review, then /start-work to activate boulder execution.
SkillCFreshness: C★ 11Claude Code +9 - 0.539score
ast-grep
mihazs
Structural code search and rewrite via the bundled ast_grep MCP server.
SkillCFreshness: C★ 11Claude Code +9
Cite this page
ulw-loop skill. Agent Mart. https://agentmart.dev/skills/mihazs-ulw-loop (last updated 2026-10-24)