pharma-conference-insight-report
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data fetched from external URLs, including company websites and conference news (Steps 1, 2, and 4). This ingestion of untrusted external content represents a surface for indirect prompt injection, where malicious instructions embedded in source materials could influence the generated report.
- Ingestion points: Web search and crawling of
company_urland conference data in SKILL.md (Steps 1-4). - Capability inventory: Python script execution for file writing (Step 8) and network data retrieval.
- Boundary markers: No explicit delimiters or sanitization steps are defined for the fetched content.
- Sanitization: Not explicitly implemented in the provided SOP.
- [COMMAND_EXECUTION]: The instructions mandate the use of Python scripts to write the final HTML report to the
@session/directory (Step 8). This is implemented as a standard procedure to ensure complete file generation and avoid content truncation within the agent's environment. - [EXTERNAL_DOWNLOADS]: The skill performs targeted searches and data retrieval from pharmaceutical conference websites and corporate domains to gather competitive intelligence. These operations are consistent with the skill's stated purpose of providing industry insights.
Audit Metadata