The Agent Skills Directory

[SAFE]: The skill follows secure development practices, such as recommending yaml.safe_load for file validation and focusing on behavioral requirements rather than implementation details.
[COMMAND_EXECUTION]: The skill uses the evaluate_matlab_code tool to interact with Simulink and Requirements Toolbox APIs. This is a core part of its intended functionality for analyzing models and managing requirements artifacts.
[COMMAND_EXECUTION]: Includes a Python-based validation step for YAML files. The command uses yaml.safe_load, which prevents the execution of arbitrary code during the parsing of generated requirement artifacts.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from Simulink models to generate requirement drafts.
Ingestion points: Untrusted data enters the context via the model_read, model_query_params, and model_resolve_params tools (SKILL.md, manifest.yaml).
Boundary markers: None explicitly defined for the model data stream.
Capability inventory: Uses evaluate_matlab_code for MATLAB operations and python for YAML validation (SKILL.md).
Sanitization: No explicit sanitization of model block names or parameter values is mentioned, though the use of yaml.safe_load during validation mitigates risk from generated content.

generate-requirement-drafts