humanizer

Pass

Audited by Gen Agent Trust Hub on Nov 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and transform arbitrary user-provided text while having access to powerful file-system tools like Write and Edit. This creates a vulnerability where malicious instructions embedded in the text being 'humanized' could potentially influence the agent to perform unintended actions.
  • Ingestion points: In SKILL.md, the 'Your Task' and 'Process' sections describe a workflow where the agent reads and processes input text supplied by the user or from external documents.
  • Boundary markers: The instructions do not prescribe the use of specific delimiters (such as XML tags or unique markers) to clearly separate the user's data from the agent's instructions, nor do they include a warning to ignore embedded commands.
  • Capability inventory: According to the YAML frontmatter in SKILL.md, the skill is granted access to the Read, Write, Edit, Grep, and Glob tools, enabling it to modify the file system based on its output.
  • Sanitization: There are no explicit instructions for the agent to sanitize input or filter out potential prompt injection attempts within the content it is asked to rewrite.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 22, 2026, 01:11 AM