clinicaltrials-database
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from the ClinicalTrials.gov API, which represents a potential surface for indirect prompt injection from trial descriptions or summaries.
- Ingestion points: Study metadata, including brief summaries and eligibility criteria, are fetched from the ClinicalTrials.gov API v2 in
scripts/query_clinicaltrials.py. - Boundary markers: The implementation uses structured JSON parsing but does not include explicit delimiters or instructions for the agent to ignore potential commands embedded within the medical text.
- Capability inventory: The skill is restricted to network GET requests and local CSV generation, lacking dangerous capabilities like arbitrary command execution that could be exploited via injection.
- Sanitization: Data is handled using standard JSON decoding; no specific sanitization is performed on the natural language fields retrieved from the database.
- [EXTERNAL_DOWNLOADS]: The skill communicates with
https://clinicaltrials.gov/api/v2, which is a well-known and trusted government medical research service. This connection is used solely for the intended purpose of querying study data.
Audit Metadata