Protocol Discovery, and this site's claim about it
Among the checks a thorough agent-readiness audit runs on a codebase, one item asks whether an agent working in a repository can actually discover the reusable procedures already available to it -- not just read source code, but find the named, invokable skills a maintainer already wrote down. This site is this family's own answer to that question for its cwc-* plugin fleet: rather than leaving "what skills does this repo have" as something only visible by walking plugins/*/skills/*/SKILL.md by hand, subagentskills.com makes it a durable, queryable, machine-readable catalog -- the same move every other primitive in this family makes for its own real schema.
This framing -- describing Agent Skills discoverability as one item on a broader "Protocol Discovery" checklist -- is this site's own contribution rather than a verified citation to an external isitagentready.com page found in this repo; no such reference was located here. It is offered as a plausible, house-style extension of the family's existing discipline (grounding every claim in a real file, tagging what's verified vs. asserted), not as a fact independently confirmed against that external site.
What "ready" looks like here
- Every skill has a real
SKILL.mdat a real path -- see the catalog. - Every skill's
descriptionis copied verbatim from its own frontmatter, so the trigger phrases an agent would match on are exactly what ships, not a paraphrase. - Every skill is tagged gated/open against that plugin's real
gates.toml, so an agent (or a human) can tell at a glance which skills can act unilaterally and which need operator approval.
Grounding for the underlying Agent Skills mechanism: this site's own concept page, sourced from code.claude.com/docs/en/skills.md.