The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a bundled Python script (assess-sdd-state.py) to automate the assessment of the workspace state. It also instructs the agent to utilize standard developer tools such as git and project-specific test runners (e.g., pytest, npm test, cargo test) during implementation and validation. These are intended behaviors for a development workflow tool.
[SAFE]: The skill includes explicit security and privacy constraints. It instructs the agent to identify and redact sensitive information like API keys, tokens, and passwords before committing proof artifacts to the repository.
[SAFE]: The skill implements a 'disable-model-invocation: true' configuration, ensuring that the workflow only runs when explicitly requested by the user, which prevents unauthorized autonomous execution.
[SAFE]: The logic for phase routing is based on the presence and content of files in a specific ./docs/specs/ directory, providing a predictable and transparent state machine for the workflow.

sdd