math-paper-collaboration

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [SAFE]: The skill references well-known academic and version control services for collaboration.
  • Evidence: Guides users to utilize Overleaf and GitHub for document hosting and repository synchronization (SKILL.md).
  • [NO_CODE]: The skill consists exclusively of markdown-based instructions and does not include any executable scripts or binary files.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external academic data, creating a potential surface for indirect prompt injection.
  • Ingestion points: Processes paper titles, co-author names, and BibTeX citation databases provided from external sources (SKILL.md).
  • Boundary markers: Absent; there are no instructions for the agent to use delimiters or ignore instructions embedded in the ingested academic data.
  • Capability inventory: The agent is instructed to perform file system operations to create repository structures and submission packages (SKILL.md).
  • Sanitization: Absent; no validation or sanitization steps are defined for the bibliographic content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 03:00 AM
Security Audit — agent-trust-hub — math-paper-collaboration