The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs automated build, lint, and test operations during its QA phase (Phase 3). These commands are executed locally to verify the integrity of the generated code.- [DATA_EXPOSURE]: The skill manages internal state, specifications, and implementation plans within a local .omc/ directory. This is used for session persistence and progress tracking.- [DYNAMIC_EXECUTION]: The skill autonomously generates source code based on user prompts and subsequently executes it through build and test suites as part of its verification lifecycle.- [PROMPT_INJECTION]: As the skill processes high-level user ideas into functional code, it is subject to indirect prompt injection risks. However, the workflow includes a dedicated Phase 4 validation step where a security-reviewer sub-agent is invoked to perform vulnerability checks on the generated output before finalization.

autopilot