The Agent Skills Directory

[SAFE]: The skill serves as a redirection to a canonical source file located at a relative path within the repository (.claude/skills/build-convex/SKILL.md).
[SAFE]: No shell commands, network operations, or sensitive file accesses are performed in the analyzed content.
[SAFE]: The metadata and descriptions are consistent with the stated purpose of building the Convex project from source.
[SAFE]: The configuration in agents/openai.yaml enforces explicit invocation, which is a security-positive restriction that prevents the agent from triggering the skill unintentionally.

build-convex