nature-article-writer

Pass

Audited by Gen Agent Trust Hub on Apr 19, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it is designed to ingest and process untrusted user-supplied scientific data. \n
  • Ingestion points: User-provided study notes, draft manuscripts, and figure descriptions provided as input for writing or rewriting tasks. \n
  • Boundary markers: Absent; the skill does not implement delimiters or specific instructions to the agent to ignore embedded commands within user data. \n
  • Capability inventory: Local script execution via scripts/nature_preflight.py, file system read/write access for manuscript assets, and optional web access for journal guides. \n
  • Sanitization: Absent; there is no evidence of filtering or escaping external content before it is processed by the agent or the preflight script. \n- [SAFE]: The Python script scripts/nature_preflight.py is a benign tool that uses standard libraries and regular expressions to provide structural and stylistic feedback; it performs no network operations or sensitive file access. \n- [SAFE]: External URL references target official documentation at nature.com, which is recognized as a well-known and trusted service in the scientific publishing domain. \n- [SAFE]: No persistence mechanisms, privilege escalation attempts, or hardcoded credentials were detected within the skill's instructions or scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 19, 2026, 09:15 PM