skill-and-agent-creator

OpenClaw skill creation turns a task, scope, and triggers into a structured SKILL.md with YAML frontmatter, safety validation, and progressive disclosure. You provide the task description, and the skill generates compliant triggers, algorithms, and reference data placement.

The Board-First protocol is a multi-agent configuration method that sets up an agent's constitution and board references before operational logic. It ensures reliable team operations by including BOOTSTRAP, QA checks, and defrost recovery patterns in the generated agent instructions.

Auditing an existing SKILL.md applies surgical changes to align public and internal safety constraints while checking for missing files. The audit process includes validation, testing guidance, and structural corrections to prevent common deployment pitfalls and brittle prompts.

Structuring references and data in OpenClaw skills requires placing documentation into a references/data directory to maintain self-contained logic. This structured documentation placement ensures the skill operates reliably and satisfies discovery requirements without external dependencies.

OpenClaw skill creation supports both workflow design and role-driven agent configuration. It transforms tasks into workflow, role, or data-driven skills while generating full multi-agent setups with constitution references and safety validation.