Skillv1.0.1

ClawScan security

Case Study Factory · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

BenignMar 18, 2026, 2:03 AM

Verdict: Benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's code, instructions, and requirements are coherent with its stated purpose (turning project materials into case-study drafts); it uses only python3, does not reach out to network or request credentials, and its behaviors match the documentation.
Guidance: This skill appears to do what it says: produce structured case-study drafts from user-provided inputs and optionally audit a project directory or CSV. Before running: (1) review the bundled scripts if you want extra assurance (they are included and use only the stdlib); (2) avoid passing broad or sensitive paths (do not point --input at /, your home, or system directories) because the script will read files you give it; (3) prefer using the examples or a sanitized copy of project materials; (4) use --dry-run or redirect output to files for review before taking any external actions; (5) if you enable autonomous agent invocation, be aware the agent could choose to run the local script on files it can access — restrict the skill to explicit user invocation if that concerns you.

Purpose & Capability: okName/description, SKILL.md, README, resources, and scripts all align: the skill converts project materials into structured case-study Markdown and can audit directories or CSVs. Required binary (python3) is appropriate and no unrelated credentials, binaries, or config paths are requested.
Instruction Scope: noteSKILL.md confines the skill to producing draft outputs and using local resources. The provided script reads files and directories the user supplies and performs pattern scans (e.g., searching for headings, CSV fields, or secret-like strings). This is appropriate for a content/audit tool, but it means the tool will read any path you pass it—do not point it at broad or sensitive locations (e.g., / or your whole home) unless you intend that.
Install Mechanism: okNo install spec is present (instruction-only with bundled script). That is low-risk: nothing is downloaded or installed from remote sources and the script relies only on the Python standard library.
Credentials: okThe skill requires no environment variables or credentials. The script does not reference external auth tokens or network endpoints; it only reads local files supplied by the user.
Persistence & Privilege: okalways is false and the skill is user-invocable. It does not request permanent presence or modify other skills or system-wide settings. Autonomous invocation is allowed by platform default but not inherently risky here given the limited footprint.