Skillv1.0.2

ClawScan security

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

BenignFeb 25, 2026, 4:49 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill is an instruction-only creative synthesizer that only uses conversation context to produce a song, visual guide, and TED-style talk — its declared requirements and behavior are consistent with its stated purpose.
Guidance: This skill is coherent and appears to do what it says: synthesize creative artifacts from the conversation. Before using it, be mindful that it reads the full conversation context and the TED-talk output is intended to include concrete, specific details from that context — so do not invoke it in threads that contain secrets, proprietary code, or confidential information you wouldn't want reproduced. Consider invoking the skill explicitly (not letting the agent run it autonomously) when working with sensitive material, and review/redact outputs before sharing externally. If you need stricter guarantees, avoid providing sensitive context or disable autonomous invocation at the agent level.

Purpose & Capability: okName/description (creative synthesis: song, visual concept, TED talk) match the SKILL.md instructions. No binaries, env vars, or config paths are requested, which is proportional for an instruction-only, conversation-based creative skill. Optional component skills are listed as helpers, not required.
Instruction Scope: noteRuntime instructions direct the agent to 'read full conversation context' and to ground the TED talk in 'concrete, specific details from context provided.' This is coherent for the stated goal, but it means the skill may surface sensitive or proprietary information that exists in the conversation. The skill explicitly states it does not read workspace files or project artifacts, which matches its declared scope.
Install Mechanism: okNo install spec or code files are present; instruction-only skills are low-risk because nothing is written to disk. The README references an openclaw install command but there is no underlying installer declared here.
Credentials: okNo environment variables, credentials, or config paths are requested. This is proportionate to a purely conversational creative skill.
Persistence & Privilege: okalways: false and no special persistence or cross-skill configuration changes are requested. disable-model-invocation is false (normal); the skill can be invoked autonomously but that is the platform default and not combined with broad privileges here.