rail-debug
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions direct the agent to ingest and act upon data from external documentation files which may contain malicious instructions intended to influence agent behavior. * Ingestion points: The skill instructs the agent to read project-specific documentation such as docs/monorail/CONTEXT.md and files within docs/monorail/adr/. * Boundary markers: The instructions lack explicit boundary markers or requirements for the agent to treat documentation content as untrusted. * Capability inventory: The agent is granted shell access (ls), file system access, and the ability to execute generated test scripts and CLI commands. * Sanitization: No sanitization or validation is performed on the content ingested from the repository documentation.
- [COMMAND_EXECUTION]: The skill methodology relies on the agent generating and executing shell commands and scripts to establish reproduction loops. * Evidence: Phase 1 of the skill recommends using CLI invocations, HTTP scripts via curl, and test harnesses, directing the agent to ensure these are agent-runnable.
Audit Metadata