procore-hello-world
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves and displays data from external Procore API endpoints, which could theoretically contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: API response data in
SKILL.md(Steps 1, 2, and 3) is accessed viaprojects.json(),rfi.json(), andsubmittals.json(). - Boundary markers: Absent; the code prints the content of the API response fields (e.g.,
p['name'],s['title']) directly to the console output without delimiting instructions. - Capability inventory: The agent uses Python's
requestslibrary to perform network operations based on these instructions. - Sanitization: Absent; the skill does not filter or sanitize the strings returned by the API before they are processed by the agent.
Audit Metadata