minimize-repro-for-diagnostic-test
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection by ingesting data from untrusted external sources to influence its behavior.
- Ingestion points: The skill reads bug details from YouTrack issues, local user project files, and API descriptions from the web.
- Boundary markers: The instructions do not specify the use of delimiters or clear boundary markers to separate user/external content from the agent's instructions.
- Capability inventory: The skill possesses the capability to execute commands (running tests), read local files (e.g., AGENTS.md), and dump logs.
- Sanitization: There is no explicit requirement for the agent to sanitize or validate the external content before using it to generate tests or dumped logs.
- [COMMAND_EXECUTION]: The skill requires the agent to execute code to verify bug reproductions.
- The instructions explicitly state to "Make sure that the problem is reproduced by running a test."
- This execution follows the ingestion of potentially attacker-controlled data from YouTrack or web documentation, which could be crafted to influence the generated test code.
Audit Metadata