clinical-data-review

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill contains instructions such as 'You are in AUTONOMOUS MODE. Do NOT ask questions', which attempts to suppress interactive agent behavior and bypass standard user confirmation loops.
  • [COMMAND_EXECUTION]: The 'SELF-EVOLUTION TELEMETRY' section instructs the agent to search for paths and append metadata to a file located at '~/.claude/projects/skill-telemetry.md', performing unauthorized file system writes in a hidden configuration directory.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input through '$ARGUMENTS' and the local codebase without implementing boundary markers or sanitization.
  • Ingestion points: '$ARGUMENTS' variable and various clinical data files (schemas, models, APIs) in the working directory.
  • Boundary markers: None present in the prompt instructions to delimit user input from system instructions.
  • Capability inventory: Extensive file system read and search capabilities across the entire codebase.
  • Sanitization: No evidence of validation or sanitization for data ingested from the codebase or arguments.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 10:56 AM
Security Audit — agent-trust-hub — clinical-data-review