Back to skill
Skillv1.0.0
ClawScan security
prompt-compress · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMay 1, 2026, 7:32 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5.5
- Summary
- This skill appears to only analyze and compress prompt text, with no evidence of hidden network access, credentials, persistence, or destructive behavior.
- Guidance
- This looks safe for ordinary prompt compression use. Before using it on private or confidential prompts, remember that the compressed output may still contain sensitive content because the skill is designed to preserve the original meaning.
Review Dimensions
- Purpose & Capability
- okThe SKILL.md, reference material, and Python scripts are coherent with the stated purpose of prompt compression, keyword extraction, and token estimation.
- Instruction Scope
- okInstructions are limited to transforming user-provided prompt text and explaining compression strategies; they do not direct the agent to override user intent or take unrelated actions.
- Install Mechanism
- okThere is no install spec, no declared external binaries, no required environment variables, and the included scripts use standard-library text processing only.
- Credentials
- okThe scripts read prompt text from command-line arguments or stdin and print compression results; no artifact evidence shows file system scanning, network calls, credential use, or account access.
- Persistence & Privilege
- okNo persistence, background execution, privilege escalation, stored memory, or long-running worker behavior is shown in the provided artifacts.