Claude Revoke — Security skill for Claude Code
Audit and revoke everything Claude Code has been given on Linux: trusted folders, "always allow" rules, saved transcripts, hooks, MCP servers.
How to install Claude Revoke
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open atraxsrc/claude-revoke and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Claude Revoke does
Audit and revoke everything Claude Code has been given on Linux: trusted folders, "always allow" rules, saved transcripts, hooks, MCP servers. Quarantine instead of delete, one command to undo. Single Python file, no dependencies.
Alternatives in Security
- Gdpr Review — Audit the codebase (or a given path/feature) for GDPR compliance and produce a prioritized findings report 172 ★
- Claudit Sec — Security audit tool for Claude Desktop and Claude Code on macOS — single-command visibility into MCP servers 115 ★
- Kali Pentest — Kali Linux penetration testing skill for AI agents (Claude Code, OpenClaw, Hermes Agent) 100 ★
README
Every time you run Claude Code in a folder and click **"Always allow"**, it remembers. After a year you can have hundreds of trusted folders, auto-approve rules, extra directories and saved transcripts spread across your disk, and no easy way to see them all.
**claude-revoke** is an installer-style terminal UI in reverse. Instead of choosing what to install, you see everything Claude Code has been given, grouped into **STALE**, **RISKY** and **OK**, tick what you want gone, and apply. Nothing is deleted: removed items go to a quarantine folder, and one command undoes a run.
Features
- Full inventory: trusted folders, "always allow" rules, extra directories, session transcripts, per-project settings files, hooks and MCP servers.
- Grouped checkboxes: STALE (folder deleted), RISKY (broad rules such as
Bash(*), bypass mode, hooks, leaked secrets) and OK. Tick a whole group or one folder at a time. - Secret scan: finds AWS keys, private keys, GitHub/Slack/API tokens and
KEY=valuelines inside saved transcripts, and shows the file, line and a redacted preview of each, so you know what to rotate. Identifiers and placeholders likeyour_password_hereare filtered out. - Hardening: one tick adds deny rules fo
Related Skills
Orca Agent Cleanup
Safely audit and quarantine Orca-installed agent skills, hooks, and CLI residue on Windows, macOS, and Linux.
Cost Audit
Per-project Claude Code cost from local transcripts (token usage x API list prices). Read-only; cross-platform
Heeler Scan All
Run a complete security scan workflow with Heeler: secrets, vulnerabilities, and license checks. Use when the
Sigil
Automated security auditing CLI for AI agent code — quarantine-first workflow for repos, packages, and agent t
Axonflow Revoke Override
Revoke an active AxonFlow session override (emits an audit event)
Rig Audit
Analyze your real sessions + project folders and propose concrete, prioritized upgrades to the SGRR AGI V2 rig
Related Agents
Settings Usability QA
Deep usability review of Tulip's Settings page — whether an operator can find a control, understand what it do
Miner
Parses structured data at volume — JSONL transcripts, logs, ledgers, CSV, counters — and returns the extracted
Skills Extractor
Extract legal skills from a public marketplace into the local ./lawve/ quarantine for curator review. Initiall
