rt-ladder-ev
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a command-line interface to invoke a local Python script located at
tools/ladder_ev.py. This script performs the core probability calculations based on provided journal metrics (months to decision, acceptance rates). The execution is confined to local project tools and does not involve arbitrary or unsanitized shell execution. - [INDIRECT_PROMPT_INJECTION]: The skill processes project-specific data which serves as a potential surface for indirect injection, though its use of structured data for arithmetic purposes makes this a low-risk surface.
- Ingestion points: Data is read from the
paper-profile.ymlfile, specifically theambition,constraints.clock, andhistoryfields. - Boundary markers: The instructions do not specify explicit delimiters or "ignore" instructions for the ingested YAML content.
- Capability inventory: The skill has the capability to execute the
tools/ladder_ev.pyscript via a shell command. - Sanitization: The skill instructions do not detail specific sanitization for the input YAML, but the inputs are treated as numerical parameters for a calculation script.
Audit Metadata