content-humanizer

Pass

Audited by Gen Agent Trust Hub on Apr 5, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process arbitrary user-supplied text for transformation without utilizing boundary markers or explicit isolation instructions.
  • Ingestion points: The skill explicitly instructs the agent to read and humanize content pasted by the user.
  • Boundary markers: There are no instructions for the agent to use delimiters (e.g., XML tags or triple backticks) to separate the untrusted user input from its own operating instructions.
  • Capability inventory: The skill involves complex text manipulation and references the execution of a local Python script (scripts/humanizer_scorer.py).
  • Sanitization: No sanitization or validation logic is defined to check for embedded instructions within the user-provided drafts.
  • [COMMAND_EXECUTION]: The skill instructions mention generating a "Humanity score" by running a local script at scripts/humanizer_scorer.py. This script is missing from the provided file set, which prevents verification of its behavior and security posture.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 5, 2026, 08:06 PM
Security Audit — agent-trust-hub — content-humanizer