reproduce-software-bugs
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The utility script
assets/python-delimiter-repro/verify.pyexecutes a bundled bug reproduction script (repro.py) usingsubprocess.run. This execution is performed using a fixed local path to verify the expected software failure behavior. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data such as bug reports, log files, and source code, which creates a potential surface for indirect prompt injection.
- Ingestion points: External bug reports, pull request bodies, and logs are identified as primary data sources in
SKILL.mdandreferences/enterprise-operation.md. - Boundary markers: Documentation in
references/enterprise-operation.mdexplicitly instructs the agent to treat these sources as untrusted and ignore any embedded instructions. - Capability inventory: The skill facilitates command execution through standard development tools and includes a local verification script (
verify.py). - Sanitization: The instructions recommend manual sanitization and synthesis of data to preserve privacy while maintaining the bug's failure signature.
Audit Metadata