My Humanizer
PassAudited by ClawScan on May 1, 2026.
Overview
This appears to be a local writing-rewrite helper with no credential use or destructive code, though its package identity metadata is inconsistent and optional verification should be kept to metrics only.
Before installing, confirm you trust the publisher and that the My Humanizer/verified-humanizer metadata mismatch is expected. For normal use, the artifacts show a local rewriting workflow; if you enable verification, keep it to non-sensitive metrics and do not send original or rewritten private text.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Users may have a harder time confirming that the installed package is the exact skill and publisher they intended.
The supplied registry metadata identifies the evaluated skill as My Humanizer / my-humanizer version 1.0.1 with a different owner ID, while the packaged metadata and SKILL.md identify verified-humanizer; this is a provenance/name consistency issue, not evidence of unsafe behavior.
"slug": "verified-humanizer", "version": "0.0.7"
Verify the publisher, slug, and version before installing, especially because the source and homepage are not provided.
When active, the skill may guide the agent’s behavior toward the humanizer workflow, but the shown guidance does not force unrelated actions.
The hook injects a system-role bootstrap reminder. The instruction is purpose-aligned and privacy-preserving, but users should know the skill can add session guidance.
return { role: "system", content: "[Verified Humanizer] Rewrite locally, measure before/after changes..." }Review and accept the hook behavior if you want the agent to receive this reminder during relevant sessions.
If verification is requested, derived evaluation metrics could be shared with a verifier, but the artifacts do not show raw text transmission.
The skill allows optional verification, which may cross a data boundary, but it explicitly limits this to metrics and forbids sending original or rewritten text.
If verification is used, send only structured evaluation data
Use verification only when needed, confirm what verifier is being used, and ensure no private text or secrets are included.