atomic-decomposition

Pass

Audited by Gen Agent Trust Hub on Feb 22, 2026

Risk Level: SAFE
Full Analysis
  • Indirect Prompt Injection (LOW): The skill possesses a surface for indirect prompt injection because it processes untrusted research ideas and paper descriptions provided by users or external sources.
  • Ingestion points: The $0 argument in SKILL.md and placeholders like {idea_description} and {atom_name} in references/decomposition-prompts.md allow arbitrary text input.
  • Boundary markers: Absent. The prompts do not use specific delimiters or instructions to ignore embedded commands within the research material.
  • Capability inventory: The skill is limited to information extraction, search (implied), and formatting. It does not contain capabilities for file writing, network exfiltration, or code execution.
  • Sanitization: Absent. There is no evidence of input validation or escaping for the research content.
  • Risk Assessment: Since the skill only generates a JSON knowledge base and does not execute the code it finds, the risk is minimal and inherent to the nature of LLM processing of external data.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 22, 2026, 05:00 AM