Back to skill
Skillv2.2.0
ClawScan security
Humanizer · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 28, 2026, 3:53 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill's instructions, files, and requested capabilities are consistent with a text-editing 'humanizer' and do not request unrelated credentials, installs, or network access.
- Guidance
- This skill appears internally consistent for editing text: it asks the agent to detect AI patterns and rewrite prose, and it declares no network or credential access. Two practical cautions: (1) The skill's frontmatter allows file operations (Read/Write/Edit/Grep/Glob). Those are reasonable for a document editor, but if you want to restrict file enumeration or bulk searches, consider limiting those permissions before enabling the skill. (2) The instructions encourage injecting 'personality' and opinions to make text sound human — review outputs for unwanted assertions or hallucinated facts, especially if the original text included sensitive or factual content. If either of these concerns matter, ask the developer to remove Grep/Glob from allowed-tools or to clarify scope limits.
Review Dimensions
- Purpose & Capability
- okName/description (remove AI patterns, make prose more human) lines up with the included SKILL.md, README, tests, and changelog. It does not request credentials, binaries, or unrelated services.
- Instruction Scope
- noteThe runtime instructions focus on scanning and rewriting text and explicitly aim to preserve meaning while injecting 'voice'. That can legitimately involve adding new phrasing or subjective tone (which may introduce stylistic judgment or mild hallucination), but the SKILL.md does not instruct the agent to read arbitrary system files or transmit data externally. The frontmatter does include file-related tools (Read/Write/Edit/Grep/Glob), which are reasonable for document editing but broaden what the agent could access if the platform grants those permissions.
- Install Mechanism
- okInstruction-only skill with no install spec and no code files to execute. Lowest-risk install footprint.
- Credentials
- okNo environment variables, credentials, or config paths are required. The skill does not ask for unrelated secrets or cloud keys.
- Persistence & Privilege
- okalways is false and there is no evidence of the skill modifying system-wide settings or other skills. Autonomous invocation is enabled by default (normal) but not combined with other risky requests.