medical-chart-review

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because its primary function is to process and summarize clinical data that could contain malicious instructions.\n
  • Ingestion points: Processes clinical text from EHR systems, progress notes, and discharge summaries as described in SKILL.md and README.md.\n
  • Boundary markers: Includes a mandatory Safety & Compliance Gate in SKILL.md §0 that requires the agent to verify HIPAA status and review scope before processing data.\n
  • Capability inventory: Performs clinical summarization and data abstraction based on predefined templates; no shell execution or external network operations are configured in the skill instructions.\n
  • Sanitization: Lacks specific instructions for sanitizing or escaping clinical text provided by the user.\n- [EXTERNAL_DOWNLOADS]: The skill's installation documentation references fetching content from official and vendor-controlled repositories.\n
  • Evidence: README.md specifies using git clone from the vendor's GitHub repository and npx for installation.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 01:57 AM
Security Audit — agent-trust-hub — medical-chart-review