anti-hall
claudecodexVerify-first guardrails for Claude Code: hooks that block made-up APIs, force-push, AI commit credit and unverified "done" claims.
61 skills · hooks
Benchmark probes only (B0, docs/BENCHMARK-METHOD.md Amendment 3): logs every hook payload it receives to .probe/ and emits uuid tokens so a trace shows where each hook's context lands.
Not in a marketplace yet, so it loads for one session with --plugin-dir.
Plugin probe-stub in talas9/anti-hall. Commands read from the repo's manifest — 20 days ago.
This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.
node "${CLAUDE_PLUGIN_ROOT}/hooks/probe.js" PreToolUse
node "${CLAUDE_PLUGIN_ROOT}/hooks/probe.js" PostToolUse
node "${CLAUDE_PLUGIN_ROOT}/hooks/probe.js" Stop
node "${CLAUDE_PLUGIN_ROOT}/hooks/probe.js" SessionStart
Verify-first guardrails for Claude Code: hooks that block made-up APIs, force-push, AI commit credit and unverified "done" claims.
61 skills · hooks
Benchmark reference only: a degenerate guard that rejects every Bash, Edit, Write, MultiEdit and NotebookEdit call.
hooks
Benchmark reference only: loads nothing (no hooks, skills or agents). Stands in for the no-plugin arm when arms run as separate, interleaved invocations.
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.