rubric-writer

Pass

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: LOWPROMPT_INJECTIONNO_CODE
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill possesses an indirect prompt injection surface because it processes untrusted data from academic artifacts to generate a review.
  • Ingestion points: output/CLAIMS.md, output/MISSING_EVIDENCE.md, and output/NOVELTY_MATRIX.md.
  • Boundary markers: Absent. The instructions do not specify delimiters or tell the agent to ignore instructions embedded in the source markdown files.
  • Capability inventory: None. The skill does not use subprocess, exec, or network libraries. It only outputs a text-based REVIEW.md file.
  • Sanitization: Absent. The inputs are used directly for reasoning without escaping or validation.
  • Risk Assessment: The severity is low because the skill lacks the capabilities to perform exfiltration, file system modification (outside the output path), or command execution.
  • [NO_CODE] (SAFE): The skill contains no executable scripts (Python, JS, Shell). It consists entirely of Markdown and YAML frontmatter, significantly reducing the attack surface for traditional exploits.
Audit Metadata
Risk Level
LOW
Analyzed
Feb 16, 2026, 12:33 AM