Omg Verifier — Development agent for Claude Code
Isolated executor for the /verify trust core — runs the tiered verification harness, judges by assertions (never impression), and on behavioral failure runs a bounded implementation-only self-fix loop.
How to install Omg Verifier
Installs to ~/.claude/agents/ahsan0444-jintech-omg-dev-omg-verifier.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/ahsan0444/jintech-omg-dev/HEAD/agents/omg-verifier.md -o ~/.claude/agents/ahsan0444-jintech-omg-dev-omg-verifier.md Restart Claude Code, or start a new session, for it to be picked up.
What Omg Verifier does
name: omg-verifier description: Isolated executor for the /verify trust core — runs the tiered verification harness, judges by assertions (never impression), and on behavioral failure runs a bounded implementation-only self-fix loop. Returns a STATUS schema, never prose. Use only from the /verify skill. model: sonnet tools: Read, Edit, Bash, Grep, Glob, ToolSearch, mcp__plugin_jintech-omg-dev_code-review-graph__query_graph_tool, mcp__plugin_jintech-omg-dev_code-review-graph__semantic_search_
Alternatives in Development
- Nelson — Coordinates multi-agent work using Royal Navy squadron patterns with sailing orders, risk-tiered quality gates 414 ★
- Sanitizer — Generic sanitizer worker agent 262 ★
- Empathy Designer — 社交货币与共情设计师。根据大纲和伤疤细节,设计文章的分享动因(Impression Management),建立 Share Map 187 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Omg Code Simplifier
Bounded write-heavy simplifier. Small diffs only; cannot self-approve.
Voice Agent
Judges dialogue and player-facing lines against the owning title's voice bible — per-line verdicts with the vi
Overnight Preparer
Use this agent only from the /oss-overnight command. It prepares a fix branch for one PR in an isolated local
Memory Master
Use when a repeated failure, user correction, or WF recovery loop triggers memory writing; also use during WF
Hcs Eval Reviewer
Reviews regression-trap quality and eval harness coverage for HCS. Ensures traps capture real failure classes
Agent Tuner
Reflexion-style agent tuner. Given evidence of a recurring failure by a specific agent, proposes the smallest
Related Skills
Mad Max
Unattended backlog-drain mode — runs executor → verifier → code-reviewer loops on all ready tasks, logs stuck/
Scvd Store Execution Contract
A behavioral contract for autonomous work. Name the success criteria before acting, spend the cheapest evidenc
Clawd Agent OS
Orchestrate 50+ Claude agents in tiered hierarchies: git-worktree-isolated fleet spawning, a durable agent bus