ocaml
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted OCaml source code and project metadata (dune, opam files) provided by users. This creates a surface for indirect prompt injection attacks, as the instructions lack explicit guidance for the agent to use boundary markers or perform sanitization on this external content before processing it.
- Ingestion points: User-provided OCaml source files (.ml, .mli), compiler error logs, and project configuration files (dune, dune-project, opam) referenced in SKILL.md.
- Boundary markers: None defined in the skill instructions.
- Capability inventory: The skill instructs the agent to execute various build and test commands (dune, opam, ocamlc) and perform file system operations like project scaffolding (dune init).
- Sanitization: No specific sanitization or validation of user-provided code is required by the instructions.
- [COMMAND_EXECUTION]: The skill uses several command-line tools for OCaml development, such as dune, opam, and ocamlc, and specialized analysis tools like camelot and dead_code_analyzer. These commands are executed based on the user's project context.
- [DYNAMIC_EXECUTION]: The skill facilitates the compilation of OCaml code into executables and their subsequent execution (e.g., via dune exec), which is a core function of the development environment it supports.
Audit Metadata