Run Issues Verify Gate — Development agent for Claude Code
Adversarial verify gate for one /run-issues issue — drives the acceptance path in the running app and rejects on observed behaviour.
How to install Run Issues Verify Gate
Installs to ~/.claude/agents/roufmlp-agent-skills-run-issues-verify-gate.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/roufmlp/agent-skills/HEAD/agents/run-issues-verify-gate.md -o ~/.claude/agents/roufmlp-agent-skills-run-issues-verify-gate.md Restart Claude Code, or start a new session, for it to be picked up.
What Run Issues Verify Gate does
name: run-issues-verify-gate description: Adversarial verify gate for one /run-issues issue — drives the acceptance path in the running app and rejects on observed behaviour. Touches no code. model: inherit effort: high color: yellow
You are an adversarial VERIFY GATE for one issue. Your job is not to tick a checklist — it is to catch behaviour that technically passes while being subtly wrong.
**Orient, don't explore.** Read `primer.md` and the issue file. Nothing else unless they poin
Alternatives in Development
- Gsd Dom Verifier — Verifies live-DOM acceptance criteria for a completed execution wave using a browser MCP server 8.8k ★
- Civitai Perf Review — Reviews a feature segment in the main Civitai Next.js app (src/) for production performance — N+1 queries, uni 7.2k ★
- Verdict Reviewer — Adversarial final reviewer for a completed work order or diff — checks the change against its acceptance crite 1.4k ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Run Issues Review Gate
Adversarial review gate for one /run-issues issue — reads the diff and tries to refute the implementation. Tou
Sub Agent Tester
PROACTIVELY used to verify that an implementation actually satisfies the Issue's acceptance criteria at runtim
Run Issues Finale
Coherence finale for a /run-issues run — reviews the entire branch diff as one change, drives cross-issue seam
Ship Critic
MUST BE USED before any App Store submission. Cynical adversarial reviewer that refuses "ready" claims unless
Odd Worker
Bounded ODD writer. Use for one implementation unit that touches two or more non-trivial files, or reading tha
Po Evidence
Evidence seat on the Atlas PO Council. Rejects unobserved problem framing, separates phenomenon from workflow
Related Skills
Ship Issue
Take one or more GitHub issues/stories from open → reviewed PR (→ merged, opt-in) autonomously — worktree, sub
Cleaning Pass
Stage 2 cleaning pass - watermark/artifact removal via the detect -> gate -> mask -> LaMa inpaint -> verify lo
Skill Frontmatter Behaviour Validation
This command starts the structured process of validating and recording the observed behaviours of skill's fron