Best practices for building accurate, triggerable, and safe DragonClaw skills.
Skill development is the fastest way to expand DragonClaw.
Use clear titles and descriptions
Describe exact trigger intents
Include precise API inputs and outputs
Show concrete tool call examples
Define output formatting rules
Declare auth needs explicitly
Prefer read-only APIs by default
Require confirmation for risky actions
Keep shell usage narrow and well-defined
Avoid ambiguous triggers that match unrelated prompts
DragonClaw skills use the same SKILL.md format as OpenClaw.
SKILL.md
That means many community skills can be reused with little or no change.
Last updated 2 hours ago