At every eligible `agent/pre-step`, the plugin calls `ctx.skills.snapshot()` for the calling session's cwd, forwards the pre-step abort signal to discovery, applies exact `skill` tool visibility, and renders the ordered `name` and `description` entries. When no prior catalog exists and that view is non-empty, it adds an initial durable user-role `<system-reminder>` to a downstream `enter` decision. Catalog messages contain only those summaries; skill bodies, paths, sources, providers, and `whenToUse` hints remain outside the catalog.
Every catalog message carries the `skill-catalog` source: a `catalog`-form context whose `entries` record exactly the `name` and `description` pairs it published, plus `update` on a replacement. The digest covers those durable entries, not the rendered prose, so the surrounding `<system-reminder>` framing cannot decide whether a republish is needed and consumers never re-parse the `<available_skills>` block. The plugin scans durable session events backwards without copying them and derives the comparison baseline from the newest visible `skill-catalog` message it can read; unreadable and foreign records are skipped. When the digest changes, the downstream `enter` decision receives a durable user-role message containing the complete replacement catalog; an empty replacement explicitly retires earlier names. If no catalog remains visible but a recognizable historical catalog exists, compaction hid it and the next complete observation re-establishes the current catalog. An incomplete provider snapshot emits nothing and preserves the last-good model view for retry at the next pre-step. If no prior catalog exists and the current view is empty, no tombstone is necessary.
The catalog is omitted when no model-invocable skills are initially available, and also when that agent's tool view restricts away the shipped `skill` tool or resolves a same-name scoped shadow instead. Identity is compared against the definition this plugin registered rather than a lookup of its own name, so the plugin works mounted globally or inside one agent's composition, where `register()` files into that agent's layer alone. Visibility changes participate in the digest, keeping prompt guidance, model-visible schema, and executable dispatch aligned.
`catalogDescriptionMaxLength` controls normalized catalog descriptions; rendering XML-escapes them. Its default is `500` and values must be integers of at least `3`, which reserves room for a truncation ellipsis. The [skill catalog hot-refresh Agent Note](../../../.agents/notes/implemented/feature/2026-07-27-skill-catalog-hot-refresh.md) owns the durable initial catalog and replacement lifecycle.
Execution uses the calling agent's `session.header.cwd` so workspace-sensitive providers resolve the winning skill. A successful call returns canonical `{ name, provider, resourceBase?, content }`, excluding catalog ranking and provider-internal machinery; its Native renderer produces one text result containing `<skill_content name="...">`, `<skill_resources>`, and `<skill_instructions>`.
Resource guidance resolves only paths or URLs explicitly referenced by the instructions against `resourceBase`; scripts, references, and assets load on demand, and the result does not enumerate a skill directory. Local providers may supply a directory, while remote or embedded providers may supply a URL or opaque loading guidance.
An unresolved name reports that the skill is unknown or no longer available. Invalid names and skills whose `invocation.modelInvocable` is `false` produce distinct error results. `invocation.userInvocable` does not restrict this model-facing surface.
Tool execution does not add a synthetic context message. Its freshly loaded result is already recorded as the tool result and becomes available to the next model step without duplicating the body. Only the catalog projection adds replacement summaries.
If model-invocable skills exist and this exact `skill` tool is visible, the agent receives the catalog template below as a durable user-role message before the first request, with one data-dependent entry per sorted skill. Later membership, description, or visibility changes append a complete replacement using the same `<available_skills>` envelope; deleting every skill appends an empty envelope with an explicit instruction not to use older names.
A skill is a reusable set of task-specific instructions. The following skills are available in this session:
<available_skills>
-`<name>`: <normalized-and-capped-description>
</available_skills>
If the user names a skill, or the task clearly matches a skill's description, call the `skill` tool with the exact skill name before taking task actions. Load all applicable skills, then follow their full instructions. This catalog contains summaries only; do not infer or follow a skill's instructions until it has been loaded.
Repeated input cost scales with skill count and `catalogDescriptionMaxLength`; no initial catalog tokens are sent when the list is empty or the tool is hidden or shadowed. Each actual catalog change adds one retained complete replacement message.
The initial durable catalog is appended after the existing reusable prefix. Dynamic changes are append-only history after that catalog, so earlier reusable tokens stay intact while each newly appended catalog and later turns form a new suffix. A new or resumed instance with a changed digest may affect cache reuse from the newly appended catalog position.
Fixed schema cost per request where the tool is visible.
#### KV Cache effect
Prefix-stable while the tool definition and visibility are unchanged. Shadowing, restrictions, or plugin lifecycle changes may invalidate reuse from this schema.
Invalid or stale selections return exactly `Error: invalid skill name "<name>"`, `Error: skill "<name>" is unknown or no longer available`, or `Error: skill "<name>" is not available for model invocation`. Provider-thrown lookup text is data-dependent and receives the same `Error: <message>` wrapper.
- **The catalog omits `whenToUse`, source, and provider metadata** — routing is based only on name and a capped description; `whenToUse` remains provider metadata and is not rendered by the loaded wrapper either.
- **Loaded instruction bodies have no size cap** — a provider can return a skill large enough to consume substantial next-step context; only catalog descriptions are truncated.
- **Resources are guidance, not attachments** — the tool reports a base directory/URL/opaque hint but neither enumerates nor fetches referenced files for the model.
- **Loading is one-shot text** — there is no partial, streaming, or cached-content handle when a remote provider is slow or a skill body is large.
- **Catalog replacement is whole-list** — one changed name or description appends every currently visible summary; this keeps stale-name retirement explicit but costs tokens proportional to the catalog.
- **Bodies are not versioned** — body-only edits do not change the catalog digest or notify the model; a later tool call reads the current provider content while earlier tool results remain historical facts.
