Skillv2.1.1

ClawScan security

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

BenignApr 10, 2026, 2:05 AM

Verdict: Benign
Confidence: high
Model: gpt-5-mini
Summary: This is an instruction-only roleplay skill whose behavior and requirements are consistent with its stated purpose (a catgirl RP persona) and it does not request credentials, install code, or external/network access.
Guidance: This skill is internally coherent for catgirl roleplay and has low technical risk (no installs, no credentials). Before installing: (1) understand that while 'in-character' the assistant will refuse normal technical or factual questions until you send the exit command; (2) it asks the agent to 'remember' settings — avoid storing secrets or private data via the skill; (3) the publisher/source is unknown, so prefer installing from a trusted registry (ClawHub) or review provenance before adding it to a production agent; (4) if you need safe/NSFW policy enforcement, verify your platform's content rules because the skill includes affectionate/bedtime roleplay triggers.
Findings: [regex_scan_no_findings] expected: The static scanner found nothing — expected because this is an instruction-only skill with no code files to analyze.

Purpose & Capability: okName, description, README and SKILL.md all describe a catgirl roleplay persona and the skill does not request unrelated binaries, env vars, or config paths — requirements match purpose.
Instruction Scope: noteSKILL.md gives detailed runtime instructions (stay in-character, mode-switching, canned responses). This is coherent for a persona skill, but note it explicitly tells the agent to refuse non-roleplay questions and to 'remember' user-provided settings — that will suppress normal assistant behavior while active and could cause useful technical queries to be ignored until exit.
Install Mechanism: okNo install spec and no code files are present (instruction-only), so nothing is written to disk or fetched at install time — lowest install risk.
Credentials: okThe skill requests no environment variables, credentials, or config paths; the declared scope is minimal and proportionate to a purely conversational roleplay skill.
Persistence & Privilege: notealways:false and autonomous invocation is default. The SKILL.md asks the agent to 'remember' user-sent settings, which is consistent with conversation state but there is no declared persistent storage — be aware settings will be retained in conversation memory while active.