PreToolUse

A pre-tool hook that runs checks before the agent uses file-reading, search, or shell commands. Hooks are automatic actions triggered by specified events.

In plain words
What is it for?
Use it to apply command-level safeguards around protected folders, evaluation data, answer keys, caches, and sealed results.
Why use it?
It can block access to listed evaluation files and paths before those tools run, helping enforce the configured access restrictions.

Hook for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add hooks/brainqub3/rlm/pre-tool-use
Clone the repo
git clone --depth 1 https://github.com/brainqub3/RLM

Made for: Claude Code.

Per session not measured What this adds to a session before it is invoked.
When invoked not measured Not applicable: nothing here is loaded into a session.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Security

Grade A, and why

PreToolUse 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 2d ago.

A static scan of the commands this hook runs, not an audit. A hook is shell that executes on your machine at the event it names, which is why every command in it is printed with what was found.

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.

.claude/settings.json · 34 lines

How it starts

The opening of the file, as written. The whole thing — 34 lines — stays where its author put it; the contents beside it link to each section on GitHub.

{
  "PreToolUse": [
    {
      "matcher": "Read|Grep|Glob|Bash",
      "hooks": [
        {
          "type": "command",
          "shell": "bash",
          "command": "grep -qiE '_archive|contexts_with_labels|_cache|verify_eval|oolong_trec_coarse|eval..?readme|gold_label_stats' && printf '%s' '{\"hookSpecificOutput\":{\"hookEventName\":\"PreToolUse\",\"permissionDecision\":\"deny\",\"permissionDecisionReason\":\"Blocked: off-limits OOLONG answer key/strategy for the cold RLM eval (issue #6) - a path under _archive/, contexts_with_labels/ (incl. GOLD_LABEL_STATS.md), the manifest (oolong_trec_coarse.jsonl), the verifier/_cache, or the eval README. Matching is case-insensitive; plain prose mentions without the path are fine; score out-of-band via score.py using the default manifest path.\"}}' || true"
        }
      ]
    },
    {
      "matcher": "Read|Grep|Glob|Bash",
      "hooks": [
        {
          "type": "command",
          "shell": "bash",
          "command": "grep -qE '_sealed/' && printf '%s' '{\"hookSpecificOutput\":{\"hookEventName\":\"PreToolUse\",\"permissionDecision\":\"deny\",\"permissionDecisionReason\":\"Blocked: sealed prior-arm eval results (cross-arm leak guard, issue #6). A finished arm is moved under rlm_vs_agent_experiment/_sealed/ -- its diagnostics.json holds the gold answer and the arm answers -- so a LATER arm claude -p session cannot read it. The driver/seal_run.py move things in/out by name (path not on the command line), so this does not block them; results are unsealed back into runs/<ts>/ after all arms finish, then scored with score.py.\"}}' || true"
        }
      ]
    },
    {
      "matcher": "Read|Grep|Glob|Bash",
      "hooks": [
        {
          "type": "command",
          "shell": "bash",
          "command": "[ -n \"$RLM_CONTROL_SESSION\" ] && grep -qE 'rlm_repl|llm_query|rlm_query' && printf '%s' '{\"hookSpecificOutput\":{\"hookEventName\":\"PreToolUse\",\"permissionDecision\":\"deny\",\"permissionDecisionReaso

Read the full file on GitHub · 34 lines

Changes

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.

  1. 2d ago First seen · 34 lines scan A c0f9c5b062d1

Subscribe to this mod's changes

PreToolUse is a hook published in the GitHub repository brainqub3/RLM (393 stars, last pushed 2mo ago), licensed MIT. Its token cost is not measured: a hook is shell that never enters the context. 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 2026-08-30.