writing-shape

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from a markdown file to drive its generation process.
  • Ingestion points: The agent reads a 'raw material' file provided by the user (SKILL.md, line 10).
  • Boundary markers: The instructions lack explicit markers (like xml tags or delimiters) to separate instructions from the data in the pile, and no 'ignore embedded instructions' warnings are present.
  • Capability inventory: The skill utilizes file system read and write capabilities to process the input and output the article.
  • Sanitization: No sanitization or content validation is performed on the ingested text before the agent processes it paragraph by paragraph.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 08:22 PM
Security Audit — agent-trust-hub — writing-shape