antithesis-launch
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references the official 'snouty' repository and documentation hosted on GitHub (github.com/antithesishq/snouty). These references are used for installation guidance and project configuration templates from the vendor.
- [COMMAND_EXECUTION]: The skill performs shell command execution to manage the testing lifecycle, including 'docker compose build', 'podman compose build', 'snouty validate', and 'snouty launch'. These operations are consistent with the skill's stated purpose of automating container-based testing.
- [INDIRECT_PROMPT_INJECTION]: The skill inspects various repository files to automatically determine configuration parameters and test metadata.
- Ingestion points: Files such as 'AGENTS.md', 'README*', 'Makefile*', and the project's directory structure (e.g., 'antithesis/config').
- Boundary markers: None identified in the discovery logic.
- Capability inventory: Local file system access, container build operations via Docker/Podman, and network-based test submission via the 'snouty' CLI.
- Sanitization: No explicit sanitization of discovered repository metadata (like test names or descriptions) is mentioned before inclusion in command arguments.
Audit Metadata