Skillv1.0.0

ClawScan security

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

Scanner verdict

BenignApr 30, 2026, 12:41 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's instructions and requirements are coherent with its stated purpose (fact-checking LLM outputs against source docs/code); it reads repository files and runs local grep/git commands but requests no external credentials or installs.
Guidance: This skill will examine your repository and documentation files (it uses grep/git-style checks) to verify claims — it does not request credentials or contact external services. Before installing, ensure you are comfortable with the agent reading project files and running local commands. Also be aware the SKILL.md expects tools like git and ast-grep to be available; if they are not installed the skill may fail. If you plan to allow autonomous invocation, consider limiting the agent's workspace or running checks in a sandboxed copy of your codebase to avoid unintended exposure of sensitive files.
Findings: [no_regex_findings] expected: Scanner found no matches because this is an instruction-only skill with no code files to analyze; that's expected.

Review Dimensions

Purpose & Capability: okThe name/description (hallucination detection for RAG/code/docs) match the SKILL.md: extracting claims and checking them against code, docs, and git history is exactly what you'd expect. There are no unrelated credential or network requirements.
Instruction Scope: noteInstructions direct the agent to parse model output and run local verification commands (grep, git, ast-grep) against repo paths like src/ and docs/. This is within scope, but the guidance implies the agent will read project files — users should expect local source access. Some steps are phrased broadly (claim extraction heuristics) which gives the agent discretion about what to read, but that discretion is aligned with the skill's goal.
Install Mechanism: okNo install spec and no code files (instruction-only), which is low-risk. Note: the SKILL.md invokes tools (ast-grep, grep, git) but the skill does not declare required binaries or provide installation instructions; it may fail at runtime if those tools are absent.
Credentials: okThe skill requests no environment variables or credentials and the instructions do not ask for secrets or external tokens. All referenced operations are local repository checks, which are proportionate to fact-checking purposes.
Persistence & Privilege: okThe skill is not always-enabled and does not request persistent privileges or modify other skills/configs. Autonomous invocation is allowed (platform default) but not elevated by the skill itself.