Skillv0.1.0

ClawScan security

Runtime Doctor · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

BenignApr 14, 2026, 1:55 AM

Verdict: Benign
Confidence: high
Model: gpt-5-mini
Summary: The skill is internally consistent with its stated purpose: a small, non-destructive local diagnostic that checks for a few workspace paths and writes a JSON report; it does not access credentials or the network.
Guidance: This skill appears safe and does what it says: it checks for the existence of a few local files and writes runtime-doctor-report.json in the current working directory. Before running, consider: (1) run it in a directory where writing runtime-doctor-report.json is acceptable (it may overwrite an existing file of that name); (2) the checks include hard-coded subpaths (projects/xzenia/...), which may be irrelevant in many workspaces—this is harmless but means the report may be sparse; (3) the SKILL.md and README mention a paid 'Pro' offering and external links, but the shipped script contains no network calls or payment integration; if you later get an upgraded version, re-check the code for outbound network behavior or credential requests; (4) as a general precaution, review the small Python file yourself if you have concerns—it's short and easy to audit.

Purpose & Capability: okName/description (diagnose runtime/config drift) matches the actual code and SKILL.md: the included script checks specific workspace state paths and generates a local report. There are no unrelated dependencies, binaries, or credentials requested.
Instruction Scope: noteSKILL.md promises 'repair suggestions' and 'only proposes safe fixes unless explicitly authorized', but the included runtime_doctor.py only performs existence checks and writes a report (no repair logic). This is a minor mismatch in scope (advertised capabilities are broader than implemented) but not malicious. The script only checks the existence of a few paths and prints/writes JSON.
Install Mechanism: okNo install spec; this is an instruction-only / small bundled script. Nothing is downloaded or executed from external URLs.
Credentials: okNo required env vars, no credentials, and the script does not read environment variables. It only inspects filesystem paths relative to the current working directory.
Persistence & Privilege: okSkill is not 'always: true' and does not request persistent platform privileges. It writes a single report file in the current working directory and does not modify other skills or system-wide configuration.