Skillv1.0.0

ClawScan security

create-psychological-counselor（创造/蒸馏心理咨询师） · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 8, 2026, 10:14 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's requested tools, files, and runtime instructions are coherent with its stated purpose of building local, research‑oriented counselor digital twins and do not request unrelated credentials or suspicious installs.
Guidance: This skill appears to do what it says: fetch links or accept pasted/uploaded material, analyze it, and write local counselor files. Before using it, do not upload non‑anonymized counseling records or other people's private data unless you have explicit permission and have properly de‑identified them. Review the generated counselors/ directory and delete any files you no longer want stored. Remember the tool is explicitly not a replacement for professional mental health care — do not use it for crisis situations. If you need greater assurance, test it with purely synthetic or public materials first and inspect what files it creates and what external URLs it fetches.

Review Dimensions

Purpose & Capability: okName/description (create counselor digital twins) align with what the skill asks to do: collect user input or URLs/files, analyze material, and write counselor artifacts to a local counselors/{slug}/ directory. Declared allowed-tools (WebFetch, Bash, Read, Write, Edit) are appropriate for fetching source links and writing generated files.
Instruction Scope: noteSKILL.md instructs the agent to fetch web links, accept pasted or uploaded content, analyze materials, and write SKILL.md/counselor.md/meta.json into local counselors/{slug}/ directories. This stays within the stated purpose. It does, however, allow ingesting potentially sensitive counseling records — the docs repeatedly warn about desensitization/authorization, but the runtime gives the agent broad discretion to fetch and process arbitrary URLs and uploaded content, so users must avoid supplying private/identifiable records without consent.
Install Mechanism: okNo install spec; the skill is instruction-only with two small helper scripts included. There are no downloads from external URLs or package installs. Scripts are local, simple, and only perform safe filesystem tasks (listing, archiving, rolling back files).
Credentials: okThe skill does not request environment variables, credentials, or config paths. The README and tools mention local tokens/config files only as local storage guidance, but no secret injection is required — this is consistent and proportional to the skill's functionality.
Persistence & Privilege: okalways is false and user-invocable is true. The skill writes files under a user-level counselors/ directory and includes versioning scripts for local archival; it does not request system-wide configuration changes or other skills' credentials.