decompose-mathematical-solution-corpora

Pass

Audited by Gen Agent Trust Hub on Oct 26, 2026

Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external mathematical artifacts which may include untrusted content from research papers, notes, or code repositories.
  • Ingestion points: Artifact types such as papers, notes, Lean/formal developments, Python/native programs, and solver encodings are inventoried in SKILL.md for deep reading.
  • Boundary markers: The instructions do not specify the use of delimiters or boundary markers to isolate untrusted content from the agent's instructions.
  • Capability inventory: The skill itself does not utilize any tools for network operations, subprocess execution, or file system modifications.
  • Sanitization: There are no instructions for sanitizing or escaping the content of the analyzed artifacts.
  • [NO_CODE]: The skill consists entirely of Markdown documentation (SKILL.md and references/continuous-moves.md) and does not include any executable scripts, binaries, or automated configuration files.
  • [SAFE]: The instructions explicitly direct the agent to avoid making unauthorized changes: 'Do not file, comment, edit external systems, or make repository changes without user authorization.'
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 26, 2026, 03:39 PM