Borrowing it
Nothing to install: this file belongs to luanmorenommaciel/agentspec. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/luanmorenommaciel/agentspec/main/.claude/commands/review/judge.mdgit clone --depth 1 https://github.com/luanmorenommaciel/agentspecWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/luanmorenommaciel/agentspec/judge)<a href="https://agentmods.dev/commands/luanmorenommaciel/agentspec/judge"><img src="https://agentmods.dev/badge/commands/luanmorenommaciel/agentspec/judge/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/luanmorenommaciel/agentspec/judge"><img src="https://agentmods.dev/badge/commands/luanmorenommaciel/agentspec/judge.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00018 | $0.01034 |
| Opus 5.5 | $0.00007 | $0.00414 |
| Sonnet 5.5 | $0.00004 | $0.00207 |
| Haiku 4.5 | $0.00002 | $0.00103 |
Grade A, and why
judge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 144 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Judge Command (V0)
Get a second opinion from a non-Claude model on code or content Claude just produced.
Usage
/judge <file> # Judge a file with default model
/judge <file> --context "building DLT CDC" # Add context about the task
/judge <file> --model anthropic/claude-3.5-sonnet # Override model (still via OpenRouter)
/judge --ledger # Show today's budget usage
What This Command Does
Sends the target file + optional task context to a non-Claude model via OpenRouter. That model returns a structured verdict:
| Field | Meaning |
|---|---|
| Verdict | PASS (no high-severity issues, confidence ≥ 0.70) or FAIL |
| Confidence | 0.0 – 1.0, judge's own self-assessment |
| Summary | One-sentence gist |
| Concerns | Table of issues (severity + evidence citing line numbers or quoted strings) |
| Suggested fixes | Concrete repairs |
The verdict renders as markdown in the chat. A ledger entry goes to .claude/storage/judge-ledger.jsonl.
When to Use
Good fits:
- Schema migrations or DDL you're about to run
- IAM / RLS policies — security-sensitive
- Complex SQL that touches production data
- Terraform / CloudFormation before apply
- Any output where "confidently wrong" would be expensive
Skip it for:
- Trivial edits, renames, formatting
- Documentation prose (judge is tuned for code/config correctness)
- Anything under ~20 lines (not enough signal)
The judge is advisory. Claude is the author; you make the final call.
Setup
# One-time: get an OpenRouter key at https://openrouter.ai/keys
export OPENROUTER_API_KEY=sk-or-v1-...
# Optional: change default model (default is openai/gpt-4o-mini — cheap + capable)
export JUDGE_MODEL=openai/gpt-4o
# Optional: change daily budget ceiling (default 10 calls/day)
export JUDGE_BUDGET=25
Full setup guide: docs/getting-started/judge-setup.md
Examples
Example 1: Judge a schema migration
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 144 lines · 18 tokens per session scan A 693d0f54ce08
judge is a command published in the GitHub repository luanmorenommaciel/agentspec (256 stars, last pushed 3d ago), licensed MIT. It adds 18 tokens to every session and 1,034 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-01.
Other commands, from other repositories
refactor
Systematic code refactoring with smart scope detection.
review-branch
Review the current branch's diff against base by dispatching atomic-reviewer. No orchestration loop, no spec required — pre-flight before /commit pr or /commit merge.
init
Install the formatters this repository needs, with every command visible before it runs.
spec-trace-requirements
You are verifying that implementation changes trace back to specification requirements.
merge-conflict-analysis
You are analyzing merge conflicts for PR #${{ pr-number }}.
boot
Start a new PromptKit prompt assembly session (alias for /bootstrap).