mechanical-port-acrazie
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes source code from external repositories which serves as a potential ingestion point. Risk is addressed through mandatory fresh context isolation for reviews, explicit activation gates for all dependencies, and required user approval for writes and command execution.
- Ingestion points: Source codebase files being ported as described in SKILL.md and references/migration-pipeline.md.
- Boundary markers: Mandatory fresh context isolation for reviews, explicit activation gates for all dependencies, and required user approval for all writes and command execution.
- Capability inventory: File system writes, shell command execution (oracles and diagnostics), and subagent spawning.
- Sanitization: Reliance on an external adversarial-reviewer component and automated test oracles to validate migrated code integrity.
- [COMMAND_EXECUTION]: The workflow requires executing shell commands for test suites and diagnostics. The skill explicitly notes potential destructive effects and requires explicit user permission for all such actions.
- [DYNAMIC_EXECUTION]: The skill generates code in the target language and executes it via test oracles and diagnostic tools as part of the mechanical migration process.
- [EXTERNAL_DOWNLOADS]: The skill instructions involve verifying model support through official documentation and potentially installing necessary project dependencies, which requires explicit user authorization.
Audit Metadata