Back to skill
Skillv1.1.1
ClawScan security
pmf-strategy · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignApr 8, 2026, 3:08 PM
- Verdict
- benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- This is an instruction-only PMF guidance skill whose requirements and instructions match its stated purpose and request no credentials or installs.
- Guidance
- This skill is an instruction-only PMF adviser and appears internally consistent and low-risk: it asks for no credentials and installs no software. Before enabling, decide whether you want the agent to be able to call the skill autonomously (the platform default), and avoid pasting sensitive proprietary data into prompts when consulting any external or third-party skill. If you need stronger control, disable autonomous invocation or review the SKILL.md periodically for updates.
Review Dimensions
- Purpose & Capability
- okThe name and description (product-market fit guidance) match the SKILL.md content. The skill is purely advisory and does not request unrelated binaries, credentials, or config paths.
- Instruction Scope
- okThe SKILL.md is a guidance document telling the agent what outputs to produce (PMF assessment, validation approach, next steps). It does not instruct the agent to read system files, environment variables, or to transmit data to external endpoints outside normal agent behavior.
- Install Mechanism
- okNo install spec or code files are present. As an instruction-only skill, it writes nothing to disk and does not fetch external code.
- Credentials
- okThe skill declares no required environment variables, credentials, or config paths. There is no disproportionate request for secrets or external service access.
- Persistence & Privilege
- notealways:false (expected). disable-model-invocation:false means the agent may autonomously invoke this skill per platform defaults — this is normal for advisory skills. Because the skill requests no credentials or installs, the autonomous-invocation risk is low, but note that the agent could call the skill without an explicit user prompt if platform policies allow.