OpenClaw Permissions Audit
v1.0.0This skill should be used when the user wants to audit, review, or list the permissions and access rights held by OpenClaw. Use it for requests like "check o...
⭐ 1· 90·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name and description claim an OpenClaw permissions audit; the SKILL.md exclusively instructs reading OpenClaw config files and extracting permission-related fields. No unrelated environment variables, binaries, or installs are requested.
Instruction Scope
The SKILL.md explicitly instructs the agent to read specific files under ~/.openclaw (openclaw.json, exec-approvals.json, identity/device-auth.json, identity/device.json) and to extract narrowly defined fields. It also includes clear rules to mask tokens/private keys and avoid outputting privateKeyPem or full IDs. This scope is appropriate for an audit, but it does require the agent to access files that contain sensitive secrets (tokens, IPC auth, keys). The skill deliberately asks the agent to skip the socket field (IPC auth token) and to avoid showing secret values, which is good practice, but the agent will still have access to those secrets at runtime.
Install Mechanism
Instruction-only skill with no install spec and no code files. Nothing is written to disk or downloaded by the skill itself.
Credentials
The skill requests no environment variables or external credentials, which is proportionate. However, it requires reading multiple local config files that are likely to contain tokens, API keys, and private keys. Reading these files is expected for a permissions audit, but it is a sensitive operation and grants the agent access to secret material (even if the skill's output rules forbid printing secrets).
Persistence & Privilege
The skill does not request always:true and has no install-time persistence. It is user-invocable and uses the default model-invocation behavior. Note: any skill that can be invoked autonomously and is allowed to read local config files increases potential impact, but this skill does not request extra privileges beyond reading the listed files.
Assessment
This skill appears to do exactly what it claims: read OpenClaw config files under ~/.openclaw and produce a masked permissions report. Before installing or running it, consider the following: (1) it will read files that may contain tokens, API keys, and private keys — although the instructions forbid printing those values, the agent will have access to them during the audit; (2) only run this skill if you trust the agent and the runtime environment (an untrusted agent could ignore masking rules and exfiltrate secrets); (3) if you are unsure, inspect the local files yourself or run the skill in an isolated environment/account; (4) prefer installing only from a known publisher or adding an explicit consent step before the skill reads sensitive files. If you do not trust the skill's source, do not grant it permission to read your ~/.openclaw directory.Like a lobster shell, security has layers — review code before you run it.
latestvk977me98mbq8vv9y0x38mgyvj984gap1
License
MIT-0
Free to use, modify, and redistribute. No attribution required.