scientific-writing

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection vulnerability surface through its integrated workflows.
    • Ingestion points: It ingests external, potentially untrusted literature data via the 'research-lookup' tool to build manuscript outlines.
    • Boundary markers: No delimiters or specific protective instructions are used to separate external data from the agent's core instructions.
    • Capability inventory: The skill is granted powerful capabilities including 'Bash', 'Read', 'Write', and 'Edit' tools.
    • Sanitization: External research content is not sanitized or filtered prior to processing.
  • [COMMAND_EXECUTION]: The skill explicitly instructs the agent to invoke a local python script via the command line interface.
    • Evidence: In 'SKILL.md', the command 'python scripts/generate_schematic.py "your diagram description" -o figures/output.png' is provided to generate visual schematics, which uses the 'Bash' tool to run local files with dynamic natural language arguments.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 04:29 AM
Security Audit — agent-trust-hub — scientific-writing