@iowarp/clio-coder 0.4.6 → 0.4.8

package/docs/guide/authoring-plugins.md

@@ -0,0 +1,158 @@
+# Authoring a portable agent plugin
+
+An agent plugin packages a workflow and its supporting resources as one versioned installation. Its canonical identity is the [Agent Plugins 1.0.0 manifest](https://agent-plugins.org/schemas/1.0.0/plugin.schema.json). Clio keeps the complete package together, verifies its full-tree digest, and discovers the resources supported by the current runtime.
+
+The minimal package has a root `plugin.json` and an immediate child skill directory:
+
+```text
+lab-research/
+  plugin.json
+  skills/
+    research/
+      SKILL.md
+```
+
+```json
+{
+  "$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
+  "name": "lab-research",
+  "version": "1.0.0",
+  "description": "Research from supplied lab evidence"
+}
+```
+
+Portable identity and publisher metadata remain at the root. The root schema permits `$schema`, `name`, `version`, `description`, `author`, `homepage`, `repository`, `license`, `keywords`, and `extensions`. Names have 1 to 64 lowercase letters, digits, hyphens or periods, begin and end with an alphanumeric character, and contain neither `--` nor `..`. Clio reserves `state.json` as an installation bookkeeping name. Every package, including a local authoring install, requires an explicit Semantic Version. Numeric prerelease identifiers cannot contain leading zeros.
+
+## Native Clio components
+
+`extensions["ai.iowarp.clio"]` holds optional Clio metadata inside the portable manifest. A plugin containing `skills/<name>/SKILL.md` needs only the root identity fields above: Clio discovers the conventional `skills/` directory. A standalone skill package with a root `SKILL.md` declares `kind: "skill"`, `resources.skills: "."` and its one public skill component through the extension; the supplied skill template demonstrates that form. Clio prompts, strict agent recipes, fleets and explicit component graphs also use this metadata. The key does not select a directory: its `resources` map names the paths the package actually uses.
+
+Use `extensions["ai.iowarp.clio"]` as the manifest metadata key to declare Clio prompts, strict agent recipes, and fleets. In modern packages, resources use ordinary top-level directories (`agents/`, `prompts/`, `skills/`, `fleets/`), while `assets/` holds shared supporting files. (Historical bundles such as `materio` may nest resources under `ai.iowarp.clio/` for backward compatibility, but top-level directories are recommended for new packages.)
+
+```json
+{
+  "$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
+  "name": "lab-research",
+  "version": "1.0.0",
+  "description": "Research from supplied lab evidence",
+  "extensions": {
+    "ai.iowarp.clio": {
+      "manifestVersion": 1,
+      "compatibility": { "clio": ">=0.4.7" },
+      "resources": {
+        "skills": "skills",
+        "prompts": "prompts",
+        "agents": "agents"
+      },
+      "components": [
+        {
+          "kind": "resource",
+          "id": "methods",
+          "path": "assets/methods.md"
+        },
+        {
+          "kind": "skill",
+          "id": "research",
+          "path": "skills/research/SKILL.md",
+          "requires": ["resource:methods"]
+        },
+        {
+          "kind": "agent",
+          "id": "researcher",
+          "path": "agents/lab-research-researcher.md",
+          "requires": ["skill:research"]
+        },
+        {
+          "kind": "prompt",
+          "id": "start",
+          "path": "prompts/lab-research/start.md",
+          "requires": ["agent:researcher"]
+        }
+      ]
+    }
+  }
+}
+```
+
+Create every declared directory and file before validation. Resource kinds are `skills`, `prompts`, `agents`, `fleets`, and `themes`. Component kinds are `skill`, `prompt`, `agent`, `fleet`, `script`, `resource`, and `tool`. A component has a stable `kind:id` identity, one contained file path, and optional `requires` references to other components in the same bundle. A skill points to its `SKILL.md`. Dependencies must exist and form an acyclic graph. Native discoverable components must live beneath their declared resource root. Component metadata records relationships; it does not execute scripts or register tool implementations.
+
+Clio text can refer to `${pluginRoot}/assets/methods.md` or `${component:resource:methods}`. Prompts, agent bodies, and skills recognize prose markers with portable, space-free slash suffixes. For a filename containing spaces or Unicode, declare a component and refer to its `kind:id`. The referenced target must exist within the declaring package. Fleet argument paths use a separate complete-path parser: the reference must occupy the whole argument, and containment is checked after expanding the entire suffix, including spaces and Unicode. A reference embedded in an option such as `--file=${pluginRoot}/data.csv` is rejected; pass the path as its own argument. These are Clio reference forms; peer exports translate instructions and paths for their native host. Portable skill text should use relative file references where practical.
+
+Prompt names follow their path beneath the prompts root. The example exposes `/lab-research:start`. Give agent and fleet filenames a package prefix to keep their registered identities unambiguous. Agent recipes retain the strict version 1 contract, including required and optional tool lists, custom audience, budget, and result contract. Recipes bind skills from their own package. Interviews belong to the orchestrator; a worker returns its need for input and is redispatched after the researcher answers.
+
+## Validation and lifecycle
+
+```bash
+clio-coder library validate ./lab-research --json
+clio-coder library inspect ./lab-research --json
+clio-coder library install ./lab-research --project
+clio-coder library list --all --json
+clio-coder library drift lab-research --project
+```
+
+`library validate <package-path> --json` performs a read-only content and payload validation before any installation or registration:
+- **Manifest validity**: Parses `plugin.json` against the schema, verifying identifiers, SemVer compliance, and declared resource roots.
+- **Payload & schema validation**: Validates agent recipes against strict v1 schemas, policy budgets, and custom audience rules; parses fleet DAGs, loop boundaries, and write boundaries; and checks prompt loadability.
+- **Package reference resolution**: Ensures all `${component:...}` and `${pluginRoot}` references point to existing files within the package boundary and form an acyclic graph.
+- **Prerequisite separation**: Reports required external commands or core agents as environment prerequisites rather than failing package content validity.
+- **Read-only**: Leaves package files, repository files, and installation state roots completely unchanged.
+
+Installation validates the exact manifest bytes included in the tree digest, copies into private staging, verifies that source and staging still match, and publishes the package and installation record with rollback on failure. Pins bind filenames, entry kinds, bytes, symlink targets and empty directories. Escaping links, hard links, unsupported filesystem entries and a root `state.json` are rejected. A catalog's expected identity, version and digest cannot be bypassed with `--force`.
+
+User installations live in the Clio configuration directory's `plugins/`; project installations live in `.clio-coder/plugins/`. State is outside the bundle, in the parent `state.json`. Installations also retain their source origin for updates. A valid project copy takes precedence over the user copy; disabling that project copy suppresses both. Changed or unverifiable content cannot load or be enabled. Update refuses local drift unless explicitly forced, and replacement or removal preserves changed files in a reported recovery copy.
+
+An active session observes additions and replacements after `/library reload` or restart. Removal, disabling and digest drift revoke subsequent resource discovery even when an earlier snapshot remains committed. Executable harness tool schemas have a separate lifecycle and require a new session after changes.
+
+## Host projections and runtime capabilities
+
+Agent Plugins 1.0.0 standardizes skills and MCP declarations. Clio currently consumes portable skills and its native manifest extension; it preserves MCP files but does not execute MCP servers. Claude Code and Gemini have their own native manifests and component formats. A package exporter must translate those formats and report which features are native, adapted as instructions, or unavailable. An exported skill workflow does not establish that a peer host executes Clio fleets.
+
+The bundled `materio` plugin includes a tested Python exporter for Codex, Claude Code and Gemini. See its package README for commands and capability reports. It also provides a complete worked example of explicit component relationships and shared domain references.
+
+Two layouts need no exporter at all, because Claude Code and Codex load them directly from the unchanged package:
+
+- A standalone skill package with a root `SKILL.md`, no `skills/` directory and no top-level `skills` manifest field. Claude Code loads it as a single-skill plugin and takes the invocation name from the `SKILL.md` frontmatter. Codex finds the same directory under any of its skills roots.
+- A bundle with a real `skills/` directory. Both hosts publish exactly those skills.
+
+What a peer host does *not* load matters just as much. Claude Code scans a plugin root for `agents/`, `commands/`, `hooks/`, `output-styles/` and `.mcp.json` with no manifest asking it to. A Clio strict-v1 recipe sitting in a top-level `agents/` directory is therefore picked up and shown as a native Claude agent, tools, model and budget fields included, none of which that host honors. Clio `prompts/` and `fleets/` are inert to it.
+
+That leaves a real choice when a package should serve both audiences. Ordinary top-level `agents/` and `prompts/` directories are the recommended Clio layout and stay a valid first-class library package; the repository marketplace simply does not publish such a package, and `pnpm run library:pin` prints the reason. To publish it as well, declare its Clio resource roots at paths no peer host scans, for example `"agents": "native/agents"` in the manifest's `resources` map, and put the recipes there. The bundled `materio` package does this with its own historical directory name.
+
+Keep the skill surface intact when you edit a package: adding a `skills/` directory to a standalone skill silently stops its root `SKILL.md` from loading. `tests/extended/library-portability.test.ts` asserts all of this for every curated package, and `scripts/verify-portable-hosts.sh` re-measures it against the real host CLIs. See [coding agent interoperability](interop.md) for the exact install commands.
+
+Harness extensions register runtime tools through the [harness extension contract](harness-extensions.md). They use a separate extension manifest and installation state, and they cannot declare domain resources: prompts, agents, skills, fleets and reference files ship in a plugin. See [plugin library usage](plugins.md) for catalogs, updates, pins and terminal UI operations.
+
+## Authoring templates and contributor journey
+
+Clio Coder provides working, valid authoring templates for all five package kinds in `library/_authoring/templates/`:
+
+- `library/_authoring/templates/skill/`: Standalone skill with house conventions, triggers, and recorded eval scenarios.
+- `library/_authoring/templates/agent/`: Standalone agent recipe using strict v1 fields, `skills: []`, and `audience: custom`.
+- `library/_authoring/templates/prompt/`: Parameterized prompt with path-derived invocation (`/<name>`) and `$ARGUMENTS` expansion.
+- `library/_authoring/templates/fleet/`: Multi-agent coordination contract with a valid DAG and documented prerequisites.
+- `library/_authoring/templates/plugin/`: Composite bundle showing an agent binding a contained skill, prompts, and supporting assets.
+
+Templates are reference models and are excluded from `library/registry.yaml`. Follow this unified contributor journey:
+
+1. **Choose package kind**: Decide whether your package is a single-kind package (`skill`, `agent`, `prompt`, `fleet`) exposing exactly one public recipe, or a composite `plugin` bundle combining multiple resources.
+2. **Copy authoring template**: Copy from `library/_authoring/templates/<kind>/` into your project or repository.
+3. **Author resources**:
+   - Single-kind packages expose exactly one public resource file in their declared resource directory (`skills`, `agents`, `prompts`, or `fleets`).
+   - Keep `README.md` at the package root, outside the resource directory, so the loader does not discover it as a second public recipe.
+   - Packages may carry explicitly invoked scripts (e.g., for offline evaluation or verification tasks); these are declared metadata and are strictly distinct from harness extensions, which register runtime tools and hooks.
+4. **Validate**:
+   Run `clio-coder library validate ./lab-research --json` to verify manifest schemas, resource payloads, reference syntax, and loadability.
+5. **Pin and check**:
+   Maintainers run `pnpm run library:pin` and `pnpm run library:check` to refresh and verify full-tree distribution pins. For curated skills, maintainers also run `pnpm run skills:pin` and `pnpm run skills:check`.
+   - **Full-tree pins** (`library/registry.yaml`) cover every distributed byte to verify package integrity.
+   - **Normalized skill evidence** (`library/skills/registry.yaml`) records normalized skill body hashes as provenance evidence for reviewed instructions.
+6. **Try in isolated project**:
+   Register locally with `clio-coder library register ./lab-research --project`. Registration writes identity, version, source, and full-tree SHA-256 into the scoped `library.yaml`. Preview and install with `library install plugin:lab-research --project --dry-run`, then `library install plugin:lab-research --project`. Test invocation through `/skills`, `/run`, prompt slash commands, or `/fleet run`.
+7. **Submit a focused fork PR**:
+   Push your topic branch to your fork and submit a PR against `main`.
+
+Package dependencies live in `extensions["ai.iowarp.clio"].requires`, such as `["skill:ship"]`; copy the same requirements into the index. They name packages, while component `requires` names files within one package. Neither declares an execution hook.
+
+Declare runnable eval suites using `"evals": {"scripts": "evals/scripts.yaml"}` in the Clio extension. Paths must identify contained files. Validate with `clio-coder eval validate --package ./lab-research --eval scripts` and run with `clio-coder eval run --package ./lab-research --eval scripts`. Use the standard version-2 eval suite schema. Materio's [worked suite](../../library/plugins/materio/evals/scripts.yaml) runs Python contracts in a temporary package copy; package installation never runs evals.
+
+See [Library packages](resource-library.md) for every lifecycle command and [Library architecture](../architecture/library.md) for identity, trust, pin and namespace invariants.

package/docs/guide/built-in-agents.md

@@ -1,8 +1,5 @@
 # Clio Coder Agent Fleet
 
-> **Visual blueprint:** The source checkout includes the complete
-> [Clio Coder Agent Fleet visual reference](https://github.com/iowarp/clio-coder/blob/main/docs/html/agents_blueprint.html).
-
 Clio Coder dispatches focused fleet agents from Markdown recipes. Recipes are data files, not hidden code plugins: YAML frontmatter declares identity, tool requirements, skill bindings, audience, capability and latency classes, budget, and result contract; the Markdown body is the agent instruction text.
 
 The source of truth is `src/domains/agents/**`. Clio's agent dispatch engine and execution boundaries are built upon the [@earendil-works/pi-agent-core](https://www.npmjs.com/package/@earendil-works/pi-agent-core) library.
@@ -16,7 +13,7 @@ Clio's agent architecture distinguishes between authoring configurations and run
 *   **Recipe**: An authored Markdown file containing frontmatter configuration and an instruction body.
 *   **AgentSpec**: The normalized runtime and catalog policy object derived from a recipe.
 *   **audience**: Determines visibility and routing (`base` | `shadow` | `custom` | `internal`).
-*   **source**: Origin of the recipe (`builtin` | `extension` | `user` | `project`).
+*   **source**: Origin of the recipe (`builtin` | `plugin` | `user` | `project`).
 
 ### Discovery, Overrides, and Precedence
 At startup, Clio loads recipes in four precedence tiers:
@@ -24,19 +21,19 @@ At startup, Clio loads recipes in four precedence tiers:
 | Source | Root | Notes |
 | --- | --- | --- |
 | **Built-in** | `src/domains/agents/builtins/*.md` in the installed package | Shipped defaults. |
-| **Extension** | Each enabled extension's `agents` resource root | Loaded in stable extension-source order. An extension recipe may bind only skills from the same extension. |
+| **Plugin** | Each enabled plugin's `agents` resource root | Loaded in stable plugin-source order. A plugin recipe may bind only skills from the same plugin. |
 | **User** | `<configDir>/agents/*.md` | Per-user recipes. `<configDir>` follows Clio's XDG/platform config directory. |
 | **Project** | `.clio-coder/agents/*.md` under the current repo | Repository-local overrides and additions (custom/domain agents). |
 
 Recipe IDs are derived from filenames (e.g., `architect.md` -> `architect`). Recipes must live directly under their respective directories.
 
 *   **Customization**: User-level agents can override/customize shipped base agents.
-*   **Extension Protection**: Extension recipes cannot override any shipped builtin.
+*   **Plugin Protection**: Plugin recipes cannot override any shipped builtin.
 *   **Shadow Protection**: User or project agents can **never** override shadow or internal agents.
 *   **Built-in Protection**: Project agents cannot override any shipped built-ins; they are strictly treated as custom/domain agents.
 *   **Reserved IDs**: The IDs `worker`, `delegate`, and `auto` cannot be registered outside the builtin tier.
 *   **Local Ignored Custom Examples**: Local examples (e.g., `benchmark-runner`, `clio-dev`, `implementer`, `scientific-validator`) may exist under `.clio-coder/agents` for documentation or test purposes, but are ignored if they collide with reserved/built-in rules.
-*   **Fleet Contracts**: Shipped builtin fleet contracts (`build-test`, `build-review`, `sdlc`) live under `src/domains/agents/fleets/*.md`. Enabled-extension contracts load next, user contracts at `<configDir>/fleets/<name>.md` load after them, and project contracts at `.clio-coder/fleets/<name>.md` take highest precedence. The parser accepts contract versions 1 through 5. Version 4 introduces enforced per-step `writes` boundaries; version 5 adds plan and gate steps, per-step target or profile routes, and the `writers: 1` single-writer declaration. Deterministic code steps reference commands declared in `.clio-coder/fleets/commands.yaml`.
+*   **Fleet Contracts**: Shipped builtin fleet contracts (`build-test`, `build-review`, `sdlc`) live under `src/domains/agents/fleets/*.md`. Enabled-plugin contracts load next, user contracts at `<configDir>/fleets/<name>.md` load after them, and project contracts at `.clio-coder/fleets/<name>.md` take highest precedence. The parser accepts contract versions 1 through 5. Version 4 introduces enforced per-step `writes` boundaries; version 5 adds plan and gate steps, per-step target or profile routes, and the `writers: 1` single-writer declaration. Deterministic code steps reference commands declared in `.clio-coder/fleets/commands.yaml`.
 
 ---
 
@@ -151,7 +148,7 @@ model selection belong to dispatch, not the recipe.
 
 Every recipe must declare `name`, `description`, `budget`, and every other key in the required set; no display defaults are synthesized. `budget` must be a non-null YAML object containing `toolCalls`, `readReserve`, and `synthesis`, plus an optional `maximum` ceiling object such as architect's `maximum: {toolCalls: 150, readReserve: 16}`. The numeric fields must be safe integers, `toolCalls > 0`, and `0 <= readReserve < toolCalls`; `synthesis` must be a boolean. Unknown, missing, quoted-numeric, floating-point, null, and relationally invalid values reject the recipe with its source path and property. Scout declares `18/4/true`; Coder declares `50/5/true`. The model-visible catalog shows the declared policy, never a mutable effective cap.
 
-Only shipped recipes may declare the `base`, `shadow`, or `internal` audience. Extension, user, and project recipes must declare `audience: custom`; the discovery root determines that provenance and the parser refuses a conflicting claim.
+Only shipped recipes may declare the `base`, `shadow`, or `internal` audience. Plugin, user, and project recipes must declare `audience: custom`; the discovery root determines that provenance and the parser refuses a conflicting claim.
 
 The operator cap is independent and cannot be widened by a recipe. Dispatch clamps `toolCalls` to that cap and clamps `readReserve` to zero when canonical `read` is absent after tool admission. Reserve slots admit only `read`, not every read-class tool. Blocked non-read attempts do not consume admitted reserve slots, but they still count toward the operator attempt ceiling.
 

package/docs/guide/commands-and-modes.md

@@ -1,8 +1,5 @@
 # Commands and Modes
 
-> **Visual blueprint:** The source checkout includes the complete
-> [Commands and Modes visual reference](https://github.com/iowarp/clio-coder/blob/main/docs/html/commands_blueprint.html).
-
 Clio Coder is a terminal-first coding agent. This page is the detailed command
 reference for interactive sessions, headless runs, dispatch, verification, and
 common operator workflows; the README remains an approachable product and
@@ -73,11 +70,10 @@ For process exit codes, stdout deliverable guarantees, and machine-readable JSON
 | `clio-coder trace code-steps <rootId> [--json]` | Show the deterministic code-step records one fleet root wrote (argv, cwd, env names, exit code, duration, output digest, artifact paths). They are files beside the ledger, not rows in the mirror, so `--db` does not apply. |
 | `clio-coder trace prune [--max-age-days N] [--max-bytes N] [--db PATH] [--json]` | Apply the trace-retention policy while protecting queued and running runs; JSON reports the resolved policy, rows, runs and bytes removed, protected runs, and whether vacuum ran. |
 | `clio-coder trace sql <SELECT query> [--db PATH]` | Run one read-only query against the mirror. Only a single `SELECT` or read-only `WITH` statement is accepted; anything else exits 2. |
-| `clio-coder trace ui [--db PATH] [--port N]` | Serve the localhost-only waterfall viewer. The viewer is not part of the published package. |
 | `clio-coder dev evolve manifest init\|validate\|summarize` | Create and check typed harness change manifests. |
 | `clio-coder extensions list\|discover\|install\|enable\|disable\|remove` | Manage installed extension packages and resource roots. `clio-coder ext` is an accepted alias. |
-| `clio-coder skills list\|search\|inspect\|validate\|install\|update\|sync\|eval` | Manage discovered skills, Clio-native skills, and local marketplace installs. |
-| `clio-coder docs [topic] [--no-open]` | Serve the interactive HTML docs of a source checkout on 127.0.0.1; the npm package ships the Markdown guides only. |
+| `clio-coder library list\|search\|register\|inspect\|validate\|install\|update\|enable\|disable\|drift\|pin\|remove` | Manage packages of kind plugin, skill, agent, prompt or fleet at user/project scope; `install/update --dry-run` preview. `library skills` lists runtime skills; `library inventory --json` is the fixed GUI read. |
+| `clio-coder docs [topic] [--no-open]` | Open the documentation in the web app, rendered directly from the canonical Markdown. Reuse the configured background app or start a foreground loopback server; `--no-open` prints its launch link. |
 | `clio-coder usage report [--repo <path>] [--days <n>] [--json]` | Cross-session usage facts from session/run ledgers and retained out-of-turn calls, including known failed-compaction spending and missing coverage. The window defaults to 30 days and the JSON schema is marked experimental. |
 | `clio-coder dev share export --out <path> [--project\|--user\|--both] [--context] [--prompts] [--skills] [--settings] [--extensions]` | Export project context, prompts, skills, settings fragments, and extension bundles. |
 | `clio-coder dev share import <path> [--dry-run] [--force] [--project\|--user] [--json]` | Import a share archive with conflict reporting. |
@@ -169,8 +165,16 @@ The registry table below lists the available interactive slash commands. On a ba
 | --- | --- | --- |
 | `/quit` | `/quit` | Exit Clio Coder |
 | `/help` | `/help [query]` | Open the interactive help center showing commands and keys |
-| `/skill` | `/skill [name] [task]` | Open the Skills Hub or invoke a skill |
-| `/resources` | `/resources [skills\|prompts\|library [kind]\|extensions [reload]]` | Browse installed resources and private library entries; `extensions reload` commits a new extension generation |
+| `/skill` | `/skill <name> [task]` or `/skill off` | Invoke a skill or clear its active tool surface; `/skills` opens the browser. |
+| `/background` | `/background` | Detach the newest eligible attached dispatch through its existing owner; no eligible dispatch gives an explanatory no-op. |
+| `/interrupt` | `/interrupt <text>` | Settle the active run and send text through the interrupt owner; refusals keep its existing next-slot behavior. Missing text is correctable. |
+| `/editor` | `/editor [text]` | Edit explicit text or an empty buffer externally; return to the composer for deliberate submission. Failure preserves the command for correction. |
+| `/notifications` | `/notifications dismiss [all]` | Dismiss the oldest notice once, or explicitly dismiss all; this does not mute future notices. |
+| `/library` | `/library [inspect \| install \| remove <ref>] [import <path-or-url>] [reload]` | Open the full-screen Library with Skills, Agents, Prompts, Fleets and Plugins tabs. A named reference opens the browser on that row; `install`, `remove` and `import` show a reviewed plan first and write nothing until it is accepted. `reload` refreshes installed recipe resources. Alt+L opens the same browser. |
+| `/skills` | `/skills` | Open the Library on Skills. |
+| `/prompts` | `/prompts` | Open the Library on Prompts. |
+| `/extensions` | `/extensions [reload]` | Inspect harness extensions or reload their hooks and operator runtimes. |
+| `/interop` | `/interop` | Inspect another local coding agent and review adoption of supported resources. |
 | `/share` | `/share [runId]` | Share a worker result with the main agent |
 | `/archive` | `/archive export <path> \| /archive import [--dry-run] [--force] <path>` | Export or import a full Clio archive |
 | `/run` | `/run [--agent-profile <profile>] [--runtime <runtimeId>] [--target <id>] [--model <id>] [--thinking <level>] [--tool-profile <minimal-local\|science-local\|full-agent>] [--require <cap>] [--share] <agent> <task>` | Run a fleet agent |
@@ -178,7 +182,7 @@ The registry table below lists the available interactive slash commands. On a ba
 | `/btw` | `/btw <question>` | Ask a side question that never enters the session transcript |
 | `/oracle` | `/oracle <question>` | Ask a read-only advisor to challenge a question against this session's settled decisions |
 | `/council` | `/council [--roster <name>] [--rounds <n>] [--synthesis <judge\|vote\|none>] <task>` | Ask a roster of read-only members the same task, with an optional vote or judge synthesis |
-| `/agents` | `/agents [list\|connect]` | List native and external agents, or review detected external-agent proposals |
+| `/agents` | `/agents` | Open the Library on Agents. |
 | `/cost` | `/cost` | Show session token and cost totals |
 | `/context` | `/context compact [instructions] \| /context recall <ref> \| /context init \| /context refresh \| /context reset` | Context hub: window overlay plus compact, recall, init, refresh, and reset |
 | `/fleet` | `/fleet run [--var <key=value>] <name>` | Open Settings → Fleet, or run a fleet contract with an approval preview |
@@ -214,7 +218,7 @@ must record a passing validation receipt for every named check or a successful
 completion note alone does not satisfy acceptance. `clio-coder tasks list`,
 `hand <uN>`, `done <uN>`, and `drop <uN>` manage the same inbox as `/tasks`.
 
-Retired spellings are unrecognized commands and use the standard unknown-command diagnostic; they do not print automatic replacement hints. Use supported spellings such as `/settings targets`, `/settings chat model-picker`, and the corresponding `/resources` subcommand. Unrecognized commands are not sent to the model as prompt text.
+Retired commands are rejected before model submission. Old Library browsing forms, including `/resources`, `/library <kind>` and `/agents connect`, explain the canonical route. Browse with `/library`, `/skills`, `/agents` or `/prompts`; use `/interop` for local-agent adoption and `/extensions` for harness extensions.
 
 `/context` with no arguments opens the context-window ledger overlay, including
 the working-set section (policy, evicted items and tokens, events, recalls, churn).
@@ -314,7 +318,7 @@ The `/resume` picker accepts Page Up and Page Down to move by its 12 visible row
 Only active commands run. Typing anything command-shaped that the registry does
 not own checks the loaded prompt templates across native and foreign prompt roots.
 Built-in command names are reserved across interactive and headless modes; a
-template with the same basename is omitted from `/resources prompts` with a collision
+template with the same basename is omitted from `/prompts` with a collision
 diagnostic instead of shadowing a command on one surface and expanding on another.
 If a matching template is found in an untrusted project root, Clio prints that the
 prompt template comes from an untrusted project root and directs the operator to set
@@ -424,17 +428,19 @@ After `/resume`, Clio offers `/memory seed` when the newest handoff contains a
 structured snapshot. Seeding is explicit, deduplicated, and unavailable while
 `context.memory.enabled` is false.
 
-The `/agents connect` overlay lists the other coding agents Clio found on this machine,
-grouped `Detected`, `Configured`, and `Declined`. A detected row's detail pane
+The `/interop` overlay lists the other coding agents Clio found on this machine,
+grouped `Detected`, `Configured`, `Declined`, and `Inventory`. A detected row's detail pane
 shows the exact `integrations.externalAgents.entries` entry that connecting it would append, plus
 the two facts a new peer inherits: `projectContext: none`, so the peer receives
 the task text and never the project projection, and `toolGovernance:
 clio-coder-policy`, so its tool calls are gated by Clio safety. Press `a` to connect
-one or `d` to decline; the overlay reads the report this process already produced
-at boot and never probes on a keystroke. Accepting applies to the live session,
+one or `d` to decline. Opening the overlay refreshes the disk inventory and runs
+bounded version probes; navigation and plan approval start no agent session. Accepting applies to the live session,
 because `delegation` hot-reloads.
 
-Boot adds at most one line about interop. The current hint can still say `Run /interop to review.`, but `/interop` is unrecognized and receives the standard unknown-command diagnostic. Open `/agents connect` directly. The hint names only agents that are installed, unconfigured, and undecided, appears at most once per set of facts, and is never emitted in headless or ACP mode. Declining an agent silences it until its binary version or path changes, at which point it becomes a fresh proposal.
+Boot adds at most one line about interop. `/interop` opens the inventory and connection review. The hint names installed, unconfigured, undecided agents and stays silent in headless or ACP mode. Declining an agent silences its connection proposal until its binary version or path changes.
+
+`clio-coder interop inspect [--json]` reports host resources and wiring. `clio-coder interop adopt <host> [--kind skill|agent|prompt|plugin] [--project|--user] [--yes] [--dry-run]` shows a plan and requires approval before installing safe resources through the library. `/interop` offers the same adoption plan with source and destination scope, kind selection, and explicit approval. See [Coding agent interoperability](interop.md) for layouts, limits, and trust semantics.
 
 `clio-coder doctor` reports interop and never proposes anything. It emits one
 `ok` row per detected agent naming its version, its path, and whether it is
@@ -446,46 +452,122 @@ one, and plain `doctor` writes nothing.
 
 ## Keybindings
 
-App bindings use `Alt + <key>` as the primary scheme, plus `Shift+Tab`,
-`Ctrl+D`, `Shift+Enter` / `Ctrl+J`, and a portable `Ctrl+G` leader. Modern terminals and Linux/meta
-setups send Alt directly. Stock macOS Terminal.app needs **Use Option as Meta
-key** enabled in Settings > Profiles > Keyboard for native Alt; otherwise use
-`Ctrl+G` then the Alt binding letter.
-
-`Alt+B` and `Alt+D` are approved application-input boundary overrides of Pi's
-editor word-back and word-delete chords. Clio routes those two bindings before
-the editor sees them so the task and decision boards remain global shortcuts.
-They are intentional exceptions to the other app bindings' avoidance of Pi
-editor reserves and can be rebound through `settings.yaml.keybindings`.
+Clio has eleven direct application defaults. `/help` shows the effective keys,
+fixed leader entries, scopes and user overrides. `Ctrl+G` opens a visible action
+menu with no timeout: use a suffix or Up/Down and Enter. Unknown suffixes leave
+the menu open and preserve the draft; Ctrl+G or Esc closes it. Ctrl+C immediately
+cancels the underlying owner. The menu does not search as you type.
 
-| Binding | Action |
+| Direct default | Action | Fixed leader suffix |
+| --- | --- | --- |
+| `Alt+L` | Toggle top-level Library | `l` |
+| `Alt+M` | Toggle model picker | `m` |
+| `Alt+O` | Cycle Compact, Standard, Detailed output | `o` |
+| `Alt+U` | Toggle dashboard | `u` |
+| `Alt+W` | Toggle Workers | `w` |
+| `Alt+E` | Toggle files from Clio focus | `e` |
+| `Shift+Tab` | Cycle supported thinking effort | `t` |
+| `Ctrl+Q` | Queue draft after the whole active run; ordinary send while idle | `f` |
+| `Alt+Q` | Restore both queue kinds before the current draft, once | `q` |
+| `Ctrl+D` | Delete forward with text; exit only empty and idle with no queued messages | — |
+| `Ctrl+G` | Open or close the contextual action menu | — |
+
+Additional menu entries: `i` interrupts with the draft, `s` backgrounds the newest
+eligible attached dispatch, `g` edits the expanded draft externally, `x` dismisses
+one notification, `z` undoes the focused editable field, `r` toggles transcript
+search, `p`/`n` page the transcript and Home/End jump to its bounds. Previous/next
+prompt entries are available through arrows and Enter. `/tasks`, `/decisions`
+and `/tree` retain their boards; scoped model cycling retains configurable action
+IDs and has no default direct key.
+
+| Editing or contextual key | Behavior |
+| --- | --- |
+| `Enter` | Accept completion first, otherwise send idle input or queue at the next steering slot |
+| `Ctrl+J`, `Shift+Enter` | Composer newline; distinct Shift+Enter delivery is optional |
+| `Alt+B`/`Alt+F`, Ctrl+Left/Right, Alt+Left/Right | Word movement |
+| `Alt+D`/Alt+Delete | Delete the next word |
+| Ctrl+W, Alt+Backspace, Ctrl+Backspace | Delete the previous word; Ctrl+Backspace covers Windows Terminal's BS encoding |
+| Home/End, Ctrl+A/E | Composer line bounds, including fullscreen |
+| Ctrl+P/N | Prompt history, restoring the unfinished draft on return |
+| Ctrl+underscore | Undo; menu `z` is the fallback when the physical chord is unavailable |
+| Ctrl+R | Fullscreen transcript search; regular mode explains terminal Find |
+| Search Enter / Up | Next / previous match; Home/End edit the query |
+| Search Esc / Ctrl+C / Ctrl+R | Close search without cancelling unrelated work |
+| PageUp/PageDown, Ctrl+Up/Down | Fullscreen transcript paging / previous-next prompt; overlays own their local navigation |
+| Esc | Cancel completion or one local field/detail first; plain composer cancels bash then the active run |
+| Ctrl+C | Cancel the current search/modal/child; otherwise cancel work, clear draft, or use idle-empty double Ctrl+C to quit |
+
+Reviews, model-scope choices, permissions and interviews own their input. They
+expose no global send, navigation, background or quit actions. Top-level Library,
+model, tree and Workers scopes can expose their owner toggle; editable fields
+expose semantic undo. Nested Library reviews must finish or cancel before the
+browser toggle becomes available. Permission deletion never resolves a user
+submit binding. Enter allows only when its defined empty-draft condition holds;
+legacy LF/Ctrl+J is indistinguishable from Enter in these single-line scopes.
+
+Bracketed paste is literal and never submits, executes a slash or shell command,
+or confirms a modal. Submission needs a later deliberate key. External editing
+receives expanded paste text and preserves the draft on failure; returned text
+and recovered queues retain literal bang provenance. Input arriving during an
+asynchronous send expansion remains in the composer. Explicit release events
+are ignored before application or viewport actions. Repeats may edit, navigate
+or scroll; they cannot repeat send, toggle, cycle, confirmation, cancel or exit.
+Legacy streams cannot distinguish every held-key repeat.
+
+**Overrides and migration.** Keep overrides in `interface.keybindings`; legacy
+top-level `keybindings` remains a compatibility input. An explicit `[]` disables
+both direct access and the action's default leader entry. A default-unbound
+action may still have a leader entry. Rebinding a direct key does not change its
+fixed suffix; `leader: []` disables the menu. Unknown IDs and effective conflicts
+diagnose, and edits reload routing, components and hints together, cancelling a
+pending menu. No preferences are rewritten. Defaults move follow-up from
+Alt+Enter to Ctrl+Q and recovery from Alt+Up to Alt+Q; Alt+B/D return to editing.
+Infrequent boards use their slash commands, and interrupt/background/external
+editor/dismiss use the menu or command bridges. Explicit old choices remain.
+
+**Terminal delivery.** On stock macOS Terminal.app, Option may compose text;
+the Ctrl+G menu works without changing that setting. A temporary profile with
+Use Option as Meta enabled permits direct Alt keys. Windows Terminal/WSL2 can
+retain native Alt+Enter, Alt+arrow, clipboard, Find and zoom controls. Ctrl+J
+is the portable newline. Node raw mode normally disables flow control and
+signal generation; an SSH/mux or terminal UI can still intercept keys, so menu
+`f` is available if Ctrl+Q does not arrive. Native clipboard and terminal window
+controls remain upstream. Clio does not own keys while Yazi, an external editor
+or the host mux holds focus. Physical Windows/macOS acceptance is separate from
+Linux PTY and protocol tests; no terminal profile is installed automatically.
+
+### What Ctrl+C does
+
+`Ctrl+C` is not a single action. It resolves against the current input boundary,
+first match wins (`resolveApplicationCtrlCAction`, `src/interactive/application-controller.ts`):
+
+| State | What Ctrl+C does |
 | --- | --- |
-| `Enter` | Send draft prompt (when idle) or deliver it at the next slot of the active run (when streaming). |
-| `Shift+Enter` / `Ctrl+J` | Insert a newline into multiline editor input. |
-| `Ctrl+P` / `Ctrl+N` | Browse backward / forward through prompts accepted in this interactive process; returning past the newest entry restores the unfinished draft. |
-| `Alt+Enter` | End of turn: queue the current draft for delivery when the active run settles. |
-| `Alt+I` | Interrupt: cancel the active run and deliver the current draft now. Refused while an attached dispatch runs or a permission ask is parked; the draft then queues for the next slot. |
-| `Alt+Up` | Restore queued next-slot and end-of-turn messages to the editor. |
-| `Shift+Tab` | Cycle orchestrator thinking level (`off`, `minimal`, `low`, `medium`, `high`, `xhigh`, `max`). |
-| `Alt+T` | Open the session tree navigator (`/tree`). |
-| `Alt+U` | Toggle the footer dashboard between compact (quiet 2-zone) and expanded (4-zone urgency) layouts. |
-| `Alt+L` | Open the model and targets selector. |
-| `Alt+J` / `Alt+K` | Cycle forward / backward through the configured model set (when empty, directs the operator to `/settings chat model-picker`). |
-| `Alt+W` | Toggle the Fleet Runs board (task, run ID, live telemetry, retry, and terminal history). Inside it, `Enter` opens the selected run's live worker detail, `s` steers, and `x` cancels. |
-| `Alt+E` | Toggle the files pane docked below the session in a `--with-panes` session inside herdr; the same as `/files`. Rebind through `clio-coder.files.toggle`. |
-| `Alt+B` | Open the composite session and operator task board (`/tasks`). Approved application-boundary override of editor word-back. |
-| `Alt+D` | Open the settled interview decision board (`/decisions`). Approved application-boundary override of editor word-delete. |
-| `Alt+S` / `Ctrl+Alt+B` | Convert an active attached dispatch to a detached background batch. |
-| `Alt+O` | Cycle Output style: Compact, Standard, Detailed. Session only; save a default in `/settings interface`. |
-| `Alt+G` | Open the current input in an external editor. |
-| `Alt+X` | Dismiss footer notifications. |
-| `Ctrl+G`, then a letter | Portable leader fallback for Alt-letter actions. |
-| `Ctrl+C` | With no overlay: cancel stream, clear input, or press twice to exit. With an overlay open: close/cancel that overlay only. |
-| `Ctrl+D` | Exit when the editor is empty; otherwise delete the next character. It never exits from inside an overlay. |
-| `PageUp` / `PageDown` | In `/resume` and list overlays: scroll the selection by one full page (12 rows). |
-| `Esc` | With an overlay open: clear non-empty filter first, move up drill-down level, then close/cancel. With no overlay: cancel stream/operation or collapse dashboard. |
-
-When scripting Clio inside tmux, prefer `tmux send-keys C-m` for submit/confirm keys instead of the literal `Enter` token; some tmux/terminal combinations do not deliver `Enter` reliably.
+| An overlay owns input (transcript search, Library, Settings, model picker, a permission card) | Closes that overlay. It does not cancel a running turn and does not exit. Any armed shutdown is disarmed. |
+| Streaming or running a tool | Cancels the in-flight run. The turn seals its partial output in ledger order and the session stays open. |
+| Idle with text in the composer | Clears the draft. The press is consumed as an editor action, so it cannot become the hidden first half of an exit. |
+| Idle with messages queued | Protects the queue and says how to recover it with the bound dequeue key (`Alt+Q` by default). |
+| Idle, empty composer, nothing queued | Arms shutdown. A second `Ctrl+C` within 500 ms (`APPLICATION_DOUBLE_TAP_MS`) exits. |
+
+Closing a permission card without approving it is a decision, not a dismissal:
+the parked call is cancelled with `User cancelled this tool call from the
+permission confirmation prompt`, and a parked *worker* permission resolves as
+`deny`. The session itself keeps running.
+
+### Typing before Clio has finished starting
+
+Interactive startup mounts the editor before it loads its services, so the
+composer accepts input from the first frame. Type and press Enter during
+startup and the submission is held in a visible pending list, in order. When
+hydration completes it adopts that same editor rather than rebuilding it, and
+the held submissions drain once, in the order you sent them, through the
+ordinary slash/bash/chat pipeline. A draft you have not submitted stays in the
+composer with its cursor.
+
+If startup fails or you interrupt it before hydration finishes, the terminal is
+restored and any pending submissions and draft text are printed to stderr rather
+than discarded. `CLIO_CODER_INSTANT_SHELL=0` opts out and waits for a fully
+hydrated first frame. The seam is `src/interactive/terminal-lease.ts`.
 
 ## Live Steering
 
@@ -495,8 +577,21 @@ There are three modes, chosen per message; the default is next slot.
 | Mode | Key | Delivery |
 | --- | --- | --- |
 | Next slot | `Enter` | Between tool batches, mid-run, through `agent.steer`. The agent keeps going and reads the message before its next model call. |
-| End of turn | `Alt+Enter` | When the whole run settles and Clio would hand control back, through `agent.followUp`. A turn is the whole run, not one model round. |
-| Interrupt | `Alt+I` (or `Ctrl+G`, `i`) | Cancels the in-flight work the way `Esc` does (generation aborts; a running bash child gets SIGTERM, then SIGKILL), waits for the cancelled run to seal its tool results in ledger order, then submits the message as a fresh prompt. Anything already queued returns to the editor. |
+| End of turn | `Ctrl+Q` | When the whole run settles and Clio would hand control back, through `agent.followUp`. A turn is the whole run, not one model round. |
+| Interrupt | `Ctrl+G`, `i` or `/interrupt <text>` | Cancels the in-flight work the way `Esc` does (generation aborts; a running bash child gets SIGTERM, then SIGKILL), waits for the cancelled run to seal its tool results in ledger order, then submits the message as a fresh prompt. Anything already queued returns to the editor. |
+
+### The same three modes against a background worker
+
+The table above describes the main agent. Against a dispatched worker the three
+modes are not symmetric:
+
+- **Steer** reaches a worker. `@<agent> <text>` sends the message across the
+  worker channel; the worker acknowledges it and keeps running.
+- **Follow-up** does not. A worker has no follow-up queue, so a new goal for it
+  is a new dispatch.
+- **Interrupt** is refused while an attached dispatch is running, for the reason
+  given below. Cancel the worker itself with `Esc`, which stops the run and
+  still leaves a receipt.
 
 Interrupt is refused in two states and the message is queued for the next slot
 instead, with a notice saying why: while an attached dispatch is running (the
@@ -586,10 +681,10 @@ to execute through the existing engine worker path, the sanctioned Claude Code w
 
 | Command | Purpose |
 | --- | --- |
-| `pnpm run ci` | Local and GitHub PR gate: typecheck, lint, skills pin check, build, the deterministic test suite, and the trace-viewer suite. |
+| `pnpm run ci` | Local and GitHub PR gate: typecheck, lint, library package pin and skill audit checks, build, the deterministic test suite, and the web application suite. |
 | `pnpm run ci:release` | Maintainer release gate: `pnpm run ci`, then the `check-release` dist and packaging audit. |
 | `pnpm run typecheck` | Strict TypeScript pass. |
-| `pnpm run lint` | Biome checks plus `scripts/check-hygiene.ts`, which runs the boundary invariants, the skills pin check, and the README and docs drift rules. |
+| `pnpm run lint` | Biome checks plus `scripts/check-hygiene.ts`, which runs the boundary invariants, the library package pin and skill audit checks, and the README and docs drift rules. |
 | `pnpm test` | Focused contract and smoke files through plain `node --test`. |
 | `pnpm run build` | Production bundle through `tsup`. |
 | `pnpm run dev` | `tsup --watch`. |
@@ -774,10 +869,10 @@ The footer owns live activity. Transcript actions have static running or outcome
 
 The Clio TUI has been enhanced to maximize readability, operational focus, and command discovery:
 
-- **Adaptive Welcome Launchpad:** Before the first prompt, renders a compact launchpad with bold CAPS section tags (`WORKSPACE`, `ROUTE`, `NEXT`), honest readiness indicators, and context-sensitive next actions. Upon first prompt submission, it deliberately collapses into a single-line session header (`>C_ Clio Coder vX.Y.Z · <workspace · git branch> · <target·model · ready> · ctx ready · type a task`) so the conversation transcript owns the viewport.
-- **Unmistakable Clio Composer:** The input editor features an explicit left section tag reflecting current prompt semantics (`MESSAGE` while idle, `FOLLOW-UP` while Clio runs, and orange `STEER` when Enter steers in-flight execution). Includes the dim placeholder `Ask Clio…  / for commands` and lower-rail hint `Enter send · Shift+Enter newline` at wider widths.
+- **Welcome header:** Before your first prompt the header is three rows and no box: a masthead (`>C_ Clio Coder v0.4.7` on the left, `~/iowarp/clio-coder · main*` on the right), the route (`✓ dynamo · qwen3.8-27b`), and one next step (`describe a task · Enter to send · / for commands`). The next step names whatever actually blocks work first: no route or model sends you to `/model`, an unavailable or degraded route to `/settings targets`, a missing or stale `CLIO-CODER.md` to `/context init` or `/context refresh`. A healthy route shows no latency; a failing one shows its reason. Your first submit collapses the header to one live row, `>C_ Clio Coder v0.4.7 · dynamo · qwen3.8-27b · ~/iowarp/clio-coder · main*`, which follows a mid-session model change. `/new` returns to the three-row form; `/resume`, `/tree`, `/fork`, `/handoff` and `--continue` open collapsed. See [tui-design.md](../architecture/tui-design.md#51-welcome-launchpad--session-header) for the width-degradation order and route vocabulary.
+- **Unmistakable Clio Composer:** The input editor features an explicit left section tag reflecting current prompt semantics (`MESSAGE` while idle, `FOLLOW-UP` while Clio runs, and orange `STEER` when Enter steers in-flight execution). Includes the dim placeholder `Ask Clio…  / for commands` and lower-rail hint `Enter send · Ctrl+J newline` at wider widths.
 - **Progressively Disclosed Footer:** The compact footer uses a quiet two-zone status layout that suppresses idle decoration (`tools none`, `◌ idle`, and duplicate turn receipts). Line 1 displays workspace location, git branch/dirty state, and active phase only when meaningful; Line 2 displays the context window gauge, current Output style, and session cost. `Alt+U` toggles the expanded dashboard, which orders information by operational urgency (Activity, Context, Session, Workspace).
-- **Footer Notification Degradation Ladder:** The footer notification badge reserves the severity head (`glyph count noun`) and `[Alt+X] dismiss` tail first, allocating remaining width to an ellipsized message body. Under narrow terminal constraints, it degrades cleanly down the ladder without clipping action keys.
+- **Footer Notification Degradation Ladder:** The footer notification badge reserves the severity head (`glyph count noun`) and `[Ctrl+G x] dismiss` tail first, allocating remaining width to an ellipsized message body. Under narrow terminal constraints, it degrades cleanly down the ladder without clipping action keys.
 - **Grouped Slash Command Palette:** Typing `/` opens an autocomplete command palette grouped by operational category (`Run`, `Inspect`, `Configure`, `Sessions`) with compact argument hints. Every suggestion is the command's one canonical spelling.
 - **Voice-First Transcript & Receipts:** User (`› `) and assistant (`✦ `) prose are formatted with a two-cell hanging indent, ensuring wrapped continuation lines remain visually tied to their voice prefix. Tool ledgers maintain full terminal width. Completed turn receipts honor output verbosity (`minimal` none, `default` compact dim `turn · in N · out M`, `verbose` full receipt with call counts, cache reads/writes, reasoning provenance, and verification caveats).
 - **Transactional Settings Center:** Open `/settings` or deep-link to one of
@@ -862,3 +957,9 @@ actual behavior. Redact secrets and private repository content.
 Acceptance rows labeled Required declare expected checks and timeout limits; they are not pending executions or passing results. Inspect verification receipts for outcomes. Under high rigor, the finish gate requires the applicable passing checks or explicit limitations.
 
 Task-board guidance and ordinary continuation preserve proposal-only scope. Deferred implementation should be blocked or dropped while awaiting an explicit operator go-ahead. A skill-install decision is separate from implementation authorization, and full-auto capability does not expand the task. These are model instructions, not a guarantee of model adherence.
+
+## Package evals
+
+`clio-coder eval validate --package <path|kind:name> --eval <name>` validates a named package suite. `clio-coder eval run` with the same package flags runs it; `--user` or `--project` selects the installed copy. Materio declares `scripts` for its offline Python contracts. The experimental skill-scenario lane is `clio-coder eval skill <name|path> [--scenario <id>]`. See [Library packages](resource-library.md).
+
+The old top-level CLI `skills` and `plugins` groups are retired; use `clio-coder library`. In the editor, `/library`, `/skills`, `/agents` and `/prompts` open the shared browser. `/resources` and `/plugins` give a replacement hint. `/skill <name>` activates a skill; `/interop` handles local-agent discovery and reviewed adoption.