diagnose
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to execute arbitrary shell commands to establish reproduction loops, including the use of test frameworks, HTTP scripts via
curl, CLI tools, and headless browsers like Playwright. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a significant ingestion surface for untrusted data, as it is instructed to process stack traces, error logs, and external event payloads to build its reproduction logic. It lacks explicit boundary markers or sanitization to prevent malicious instructions embedded in these logs from influencing the agent's behavior.
- [DYNAMIC_EXECUTION]: As part of its core debugging loop, the skill generates and runs temporary code, including "throwaway harnesses" and mocked dependencies, to exercise and verify specific code paths.
- [EXTERNAL_DOWNLOADS]: The documentation provides installation instructions that fetch the skill from an external source at
emiliosheinz/agent-skills.
Audit Metadata