poly-adk-testing
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill makes extensive use of the
polyCLI to interact with the PolyAI platform and local project files. - Evidence: Commands include
poly validate,poly chat,poly test run, andpoly functions execute. - [EXTERNAL_DOWNLOADS]: The skill's metadata specifies the requirement for the official PolyAI ADK package.
- Evidence: Frontmatter indicates the tool is installed via
uv tool install polyai-adk. - [INDIRECT_PROMPT_INJECTION]: The skill identifies an attack surface where instructions could be embedded in local data files processed by the agent.
- Ingestion points: The skill reads conversation scripts from
script.txtand test cases from thetest_suite/directory. - Boundary markers: Absent; instructions do not specify the use of delimiters when reading these files.
- Capability inventory: The
polyCLI enables file reading, network communication with the PolyAI platform, and project function execution. - Sanitization: Absent; the skill does not mention pre-processing or validating the content of the input files for malicious instructions.
Audit Metadata