specialist-engineering-ai-data-remediation-engineer

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data (anomalous rows) and interpolates user-provided text into a subagent prompt.
  • Ingestion points: User input is interpolated directly into the Agent tool call prompt via the {the user's question} variable.
  • Boundary markers: The skill lacks explicit delimiters (such as XML tags or triple quotes) to separate user-provided input from the system instructions.
  • Capability inventory: The skill includes the capability to generate and execute Python code (lambdas) and SQL expressions.
  • Sanitization: The skill includes 'Rule 3', which mandates a safety check ensuring generated logic starts with lambda and rejects content containing import, exec, eval, or os.
  • [COMMAND_EXECUTION]: The subagent is instructed to generate transformation logic in the form of Python lambdas or SQL and apply them to data.
  • Logic: While the skill involves dynamic code generation, it explicitly restricts execution to a local, sandboxed environment and requires strict pattern matching to prevent execution of arbitrary system commands or unauthorized library imports.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 08:49 PM
Security Audit — agent-trust-hub — specialist-engineering-ai-data-remediation-engineer