cite-style-converter

Pass

Audited by Gen Agent Trust Hub on Nov 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill's functionality is implemented in a local script (scripts/citation_formatter.py) that uses only standard library modules such as re, json, and argparse. No external dependencies, remote downloads, or unauthorized network communications are present.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted citation text and JSON data from users, creating a surface for indirect prompt injection, although this is typical for text-processing utilities.
  • Ingestion points: Citation strings passed as command-line arguments and file contents read via the --input flag in scripts/citation_formatter.py.
  • Boundary markers: None; the skill does not wrap processed input in delimiters or provide specific instructions to the model to ignore potential commands within the data.
  • Capability inventory: The script can read files from the filesystem when a path is provided via the --input flag and returns the result to the agent's context.
  • Sanitization: The input is parsed using regular expressions and json.loads(), which provides structural validation but does not sanitize for malicious instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 22, 2026, 01:11 AM