phoenix-evals-new-metric

Pass

Audited by Gen Agent Trust Hub on Mar 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill describes a legitimate development workflow for contributing to the Phoenix repository. All file paths and operations are consistent with the project's architecture.
  • [COMMAND_EXECUTION]: The skill involves running standard build and development commands such as make codegen-prompts, pnpm build, and pnpm tsx. These commands are used to compile generated code and execute local benchmarks within the project directory.
  • [PROMPT_INJECTION]: The skill includes explicit security and accuracy guidance for writing LLM prompts. It recommends using XML-style tags (e.g., <context>, <output>) to wrap user-provided data, which is a recognized mitigation against indirect prompt injection by ensuring the model can distinguish between instructions and data.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 21, 2026, 11:56 AM