Skillv1.0.0

ClawScan security

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

BenignApr 29, 2026, 4:48 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: Instruction-only skill that coherently guides the assistant to produce shorter, calibrated answers and does not request credentials, install code, or access system resources.
Guidance: This skill is coherent and low-risk because it's only a set of response-editing instructions and requests no access to your system or secrets. Before enabling broadly, test it on non-sensitive queries to ensure important caveats aren't being removed (especially for medical, legal, or safety-critical prompts). If you notice missing disclaimers or unsafe brevity, disable the skill or adjust your prompt to require explicit safety/context. Pairing with a 'clarity-first' skill as suggested (or adding explicit rules to preserve safety-critical caveats) can help maintain necessary context while reducing fluff.

Purpose & Capability: okThe name and description (response-length calibration) match the SKILL.md content. No binaries, env vars, or config paths are required and none are referenced; the requested capabilities are proportionate to the stated purpose.
Instruction Scope: noteSKILL.md contains a clear, self-contained protocol (classify request, draft then compress, density check). It directs removing preambles, restatements, and non-essential caveats — this is consistent with the goal but gives the agent discretion to drop contextual caveats. That discretion can reduce helpful safety/context in sensitive domains (medical, legal, safety-critical). The skill also instructs not to display the internal classification to users, which is reasonable for operation.
Install Mechanism: okNo install spec or code files are present; the skill is instruction-only so nothing is written to disk and there is no external download risk.
Credentials: okNo environment variables, credentials, or config paths are required. The instructions do not reference any external secrets or unrelated services.
Persistence & Privilege: okFlags show default behavior (not always-on, model invocation enabled). The skill does not request permanent presence or modify other skills or system-wide settings.