agentsclimarketplace
Sign in

Bench

Skill RaiAnsar/peerBench/skills/bench

Run peerBench review, hunt, investigate, setup, status, and reviewer-management commands from Codex. Use when the user asks for peerBench, bench review, bench hunt, bench investigate, bench status, bench setup, or bench reviewers.From its SKILL.md

Install
npx -y skills add RaiAnsar/peerBench --skill bench

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

1.9 KB, as published. Nobody here has run it

peerBench

peerBench is a read-only Grok and MiMo panel for code review and bug hunts.

Resolve the plugin root as two directories above this skill directory. Infer the requested subcommand from the user's request and $ARGUMENTS, then run exactly one matching scripts/bench-runner.mjs command from the current workspace. Do not append every invocation to a hardcoded review command.

BENCH_SESSION_ID="${CODEX_SESSION_ID:-${BENCH_SESSION_ID:-}}" node "<plugin-root>/scripts/bench-runner.mjs" <subcommand> [arguments]

<plugin-root>, <subcommand>, and [arguments] are notation. Resolve or replace them before execution; never pass those literal tokens to the shell.

Use these subcommands:

  • setup — show reviewer keys, hook health, and active reviewers.
  • status [trace-id] — show recent traces or expand one trace.
  • review --json [--strict] [--base <ref> | <range>] — bounded one-shot review of current changes or the exact diff for a positional base..head range. Provider failure is advisory unless strict; strict requires a usable non-blocking verdict from every active reviewer.
  • hunt [focus] — run a read-only multi-reviewer bug hunt.
  • investigate <problem> — run a deeper focused investigation.
  • debug <failure> — root-cause a specific error or failing behavior.
  • reviewers [names...] — show or set the active reviewer panel.
  • on [--global] / off [--global] — enable or disable gates.
  • scorecard — show reviewer performance stats.
  • health [--all] — probe configured reviewer health.

Return command output faithfully. Do not edit files unless the user explicitly asks for fixes after seeing the peerBench output.

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.