fix(coding-agent): synced v15.13 settings, help, status, and docs surfaces

docs/keybindings.md

@@ -47,3 +47,11 @@ On Windows Terminal, `Ctrl+V` may be handled by the terminal paste command befor
 Terminals that implement OSC 5522 enhanced paste can send clipboard MIME data directly to `omp`; image pastes are attached as `[Image #N]`, while text/plain paste events keep normal paste behavior. When OSC 5522 is unavailable, bracketed paste still handles text, and a pasted single image-file path is loaded as an image when the file is readable from the `omp` host.
 
 Older unqualified action names are migrated when `keybindings.yml` is loaded, but new docs and new configs should use the namespaced action IDs above. Existing `keybindings.json` files are still accepted and migrated to `keybindings.yml`; `keybindings.yaml` is also accepted.
+
+## Built-in editor chords
+
+| Key | Action |
+| --- | ------ |
+| `Enter` | Send message |
+| `Shift+Enter` / `Alt+Enter` / `Ctrl+J` | Insert a newline |
+

docs/settings.md

@@ -345,6 +345,7 @@ A value of `-1` means "use the provider/model default" — `omp` does not send t
 | `presencePenalty` | number | `-1` | Presence penalty. |
 | `repetitionPenalty` | number | `-1` | Repetition penalty. |
 | `serviceTier` | enum | `none` | `none`, `auto`, `default`, `flex`, `scale`, `priority`, `openai-only`, `claude-only`. |
+| `fastModeScope` | enum | `both` | Which provider families `/fast on` and the fast-mode toggle affect: `both`, `openai`, or `claude`. |
 | `personality` | enum | `default` | `default`, `friendly`, `pragmatic`, `none`. |
 
 ### Retry and fallback
@@ -388,6 +389,8 @@ tools:
 | `tools.approval` | record | `{}` | Per-tool policy keyed by tool name; each value is `allow`, `deny`, or `prompt`. e.g. `omp config set tools.approval '{"bash":"prompt"}'`. |
 | `tools.discoveryMode` | enum | `auto` | `auto`, `off`, `mcp-only`, `all`. Controls dynamic tool discovery. |
 | `tools.essentialOverride` | array | `[]` | Tool names kept available even when tools are narrowed. |
+| `task.eager` | enum | `default` | `default`, `preferred`, `always`; controls how strongly the agent is pushed toward subagent delegation. |
+| `todo.eager` | enum | `default` | `default`, `preferred`, `always`; controls first-turn todo-list creation guidance. |
 | `tools.maxTimeout` | number | `0` | Max tool runtime in seconds; `0` = no cap. |
 | `tools.intentTracing` | boolean | `true` | Record per-call intent strings. |
 | `tools.outputMaxColumns` | number | `768` | Per-line byte cap for streaming output; `0` disables. |
@@ -608,7 +611,9 @@ searxng:
 | `exa.enableResearcher` | boolean | `false` | Exa researcher. |
 | `exa.enableWebsets` | boolean | `false` | Exa websets. |
 | `searxng.endpoint` | string | _(unset)_ | SearXNG instance URL. |
-| `searxng.token` | string | _(unset)_ | SearXNG token; also `searxng.basicUsername`/`searxng.basicPassword`/`searxng.categories`/`searxng.language`. |
+| `searxng.categories` | string | _(unset)_ | Optional categories passed to SearXNG when an endpoint is configured. |
+| `searxng.language` | string | _(unset)_ | Optional language code passed to SearXNG when an endpoint is configured. |
+| `searxng.token` | string | _(unset)_ | SearXNG token; also supports `searxng.basicUsername`/`searxng.basicPassword`. |
 | `auth.broker.url` | string | _(unset)_ | Auth-broker URL. Overridden by `OMP_AUTH_BROKER_URL`. |
 | `auth.broker.token` | string | _(unset)_ | Auth-broker token. Overridden by `OMP_AUTH_BROKER_TOKEN`. |
 

docs/tools/irc.md

@@ -31,16 +31,16 @@
 ## Outputs
 - Single-shot `AgentToolResult`; no streaming updates.
 - `content` is one text block:
-  - `list`: `No other agents.` or `<n> peer(s):` bullets — `id [displayName · kind · status]` plus unread count, parent, and last-activity age; a footer notes that parked agents are revived automatically when messaged.
+  - `list`: `No other agents.` or `<n> peer(s):` bullets — `id [displayName · kind · status]` plus current `activity` when a running peer reports work, unread count, parent, and last-activity age; a footer notes that parked agents are revived automatically when messaged.
   - `send`: per-recipient delivery receipts (`injected` / `woken` / `revived` / `failed — <error>`); with `await: true`, the reply body or a clean no-reply timeout note.
   - `wait`: the consumed message as `[<msgId>] <from>: <body>` (with a reply-to tag), or `No message within <duration>.`
   - `inbox`: `Inbox empty.` or `<n> message(s):` bullets.
-- `details: IrcDetails`: `{ op, from?, to?, receipts?, waited?, inbox?, peers? }`. `waited` is `null` when a wait timed out; `receipts` carry `{ to, outcome, error? }`.
+- `details: IrcDetails`: `{ op, from?, to?, receipts?, waited?, inbox?, peers? }`. `peers[]` entries carry `activity?`, `unread`, `parentId?`, and `lastActivity`; `activity` is a display-only gist of the peer's current work and is cleared when the peer stops running. `waited` is `null` when a wait timed out; `receipts` carry `{ to, outcome, error? }`.
 
 ## Flow
 1. `IrcTool.createIf` constructs the tool only when `isIrcEnabled` passes and the session has both an `AgentRegistry` and `getAgentId`. There is no `irc.enabled` setting: availability is derived — true for every subagent (`taskDepth > 0`; a parent always exists) and for any session that can still spawn subagents through the task tool. Only a top-level session with task spawning unavailable has no peers, hence no irc.
 2. `execute` resolves the registry and sender id; missing either returns a text error result instead of throwing.
-3. `op: "list"`: `registry.list()` minus self and minus `aborted` agents — `parked` peers ARE listed. Each row includes the unread count from `IrcBus.unreadCount(...)` and last activity.
+3. `op: "list"`: `registry.list()` minus self and minus `aborted` agents — `parked` peers ARE listed. Each row includes `activity` for running peers, unread count from `IrcBus.unreadCount(...)`, and last activity.
 4. `op: "send"` validates `to`/`message`, rejects self-sends, and rejects `await` with `to: "all"`.
 5. Target resolution: broadcasts fan out to `registry.listVisibleTo(senderId)` (live peers only — `running`/`idle`; reviving every parked agent on a broadcast would be a stampede). Direct sends go through the bus unfiltered, so a parked recipient is revived.
 6. `IrcBus.send(...)` is fire-and-forget — it never blocks on the recipient generating anything. Delivery by recipient status:
@@ -54,7 +54,7 @@
 10. Timeouts resolve as `params.timeoutMs ?? irc.timeoutMs`, normalized: `0` disables the timeout, negative/non-finite values fall back to the default `120_000`, positive values are truncated and clamped to ≥ 1 ms.
 
 ## Modes / Variants
-- `list`: enumerate peers with status (`running`/`idle`/`parked`), unread counts, and last activity.
+- `list`: enumerate peers with status (`running`/`idle`/`parked`), current activity gist for running peers, unread counts, and last activity.
 - `send` direct: one exact peer id; wakes idle peers, revives parked ones.
 - `send` broadcast: `to: "all"` to every live peer; parked peers are skipped.
 - `send` + `await: true`: round-trip convenience — send, then wait for the next message from that peer. Marks the send `expectsReply`, enabling the busy-recipient auto-reply path when async execution is disabled.

docs/tools/learn.md

@@ -0,0 +1,43 @@
+# learn
+
+> Capture a reusable lesson to long-term memory, optionally creating or updating a managed skill in the same call.
+
+## Source
+- Entry: `packages/coding-agent/src/tools/learn.ts`
+- Model-facing prompt: `packages/coding-agent/src/prompts/tools/learn.md`
+- Key collaborators:
+  - `packages/coding-agent/src/autolearn/managed-skills.ts` — managed skill path/name validation and `SKILL.md` writes.
+  - `packages/coding-agent/src/memory-backend/local-backend.ts` — local file-backed lesson storage.
+  - `packages/coding-agent/src/hindsight/state.ts` and `packages/coding-agent/src/mnemopi/state.ts` — remote/local memory retention.
+
+## Inputs
+
+| Field | Type | Required | Description |
+|---|---|---:|---|
+| `memory` | `string` | Yes | Durable, self-contained lesson. Include what, when, and why. |
+| `context` | `string` | No | Source context for the lesson. |
+| `skill` | `{ action, name, description, body }` | No | Also create or update a managed skill. `action` is `create` or `update`; `name` is kebab-case; `body` is Markdown without frontmatter. |
+
+## Outputs
+- Memory only: `Lesson stored.` or `Lesson queued for retention.` with `details.skill = null`.
+- Memory plus skill: `<lesson result>. Created/Updated managed skill "<name>".` with `details.skill = <name>`.
+- Authored-skill name collision on `create`: error text explains that managed skills cannot override authored skills; the memory lesson was already stored or queued.
+
+## Flow
+1. `LearnTool.createIf(...)` exposes the tool only when `autolearn.enabled=true` and `memory.backend` is `hindsight`, `mnemopi`, or `local`.
+2. `execute(...)` stores the lesson through the active backend:
+   - `mnemopi` writes a fact with source `coding-agent-learn` and fails if no memory id is returned;
+   - `local` appends through the local backend and fails when sanitization leaves nothing to store;
+   - `hindsight` queues the lesson for asynchronous retention.
+3. If `skill` is present, the tool validates the managed skill name, refuses authored-skill shadowing on `create`, then writes the managed skill under `~/.omp/agent/managed-skills`.
+
+## Limits & Caps
+- Availability requires both `autolearn.enabled` and an active memory backend.
+- The skill side never edits user-authored skills; authored skills keep precedence.
+- `skill.body` must omit frontmatter; frontmatter is generated from `name` and `description`.
+
+## Errors
+- `Mnemopi backend is not initialised for this session.`
+- `Hindsight backend is not initialised for this session.`
+- `Lesson was empty after sanitization; nothing stored.`
+- Managed skill write failures surface after the memory outcome, because the lesson may already be stored or queued.

docs/tools/manage-skill.md

@@ -0,0 +1,40 @@
+# manage_skill
+
+> Create, update, or delete isolated managed skills used by the auto-learn workflow.
+
+## Source
+- Entry: `packages/coding-agent/src/tools/manage-skill.ts`
+- Model-facing prompt: `packages/coding-agent/src/prompts/tools/manage-skill.md`
+- Key collaborators:
+  - `packages/coding-agent/src/autolearn/managed-skills.ts` — managed skill directory, name validation, create/update/delete primitives.
+  - `packages/coding-agent/src/extensibility/skills.ts` — authored-skill collision checks.
+
+## Inputs
+
+| Field | Type | Required | Description |
+|---|---|---:|---|
+| `action` | `"create" | "update" | "delete"` | Yes | Operation to perform. |
+| `name` | `string` | Yes | Kebab-case skill name. |
+| `description` | `string` | Create/update | One-line description of when to use the skill. |
+| `body` | `string` | Create/update | `SKILL.md` body in Markdown, without frontmatter. |
+
+## Outputs
+- Delete: `Deleted managed skill "<name>".` with `{ action: "delete", name }`.
+- Create/update: `Created/Updated managed skill "<name>" (managed-skills/<path>).` with `{ action, name }`.
+- Authored-skill name collision on `create`: error text explains that managed skills cannot override authored skills and asks for a different name.
+
+## Flow
+1. `ManageSkillTool.createIf(...)` exposes the tool only when `autolearn.enabled=true`.
+2. Validation requires `description` and `body` for `create` and `update`; `delete` needs only `name`.
+3. `create` refuses names already claimed by authored skills, because managed skills resolve below authored skills and would never surface.
+4. Writes target `~/.omp/agent/managed-skills`; managed skills are surfaced like normal skills in future sessions.
+
+## Limits & Caps
+- This is a write-tier tool.
+- Managed skills are separate from user-authored skills.
+- `body` must omit frontmatter; it is generated from `name` and `description`.
+
+## Errors
+- Validation rejects create/update without both `description` and `body`.
+- Create fails when the managed skill already exists or an authored skill owns the name.
+- Update/delete fail when the managed skill does not exist.

docs/tools/task.md

@@ -26,9 +26,9 @@
 
 ## Inputs
 
-The wire schema is shape-swapped by `task.batch` (default on). One unit of work is the task item `{ id?, description?, assignment, isolated? }` (`isolated` only when `task.isolation.mode` is not `none`):
+The wire schema is shape-swapped by `task.batch` (default on). One unit of work is the task item `{ id?, description?, role?, assignment, isolated? }` (`isolated` only when `task.isolation.mode` is not `none`):
 
-- **Batch shape** (`task.batch` on): `{ agent, context, tasks: item[] }` — one subagent per item, all run under the same fan-out rules. `context` is **required** shared background rendered into every spawned subagent's system prompt (`CONTEXT` section); `isolated` is per item.
+- **Batch shape** (`task.batch` on): `{ agent, context, tasks: item[] }` — one subagent per item, all run under the same fan-out rules. `context` is **required** shared background rendered into every spawned subagent's system prompt (`CONTEXT` section); `role` and `isolated` are per item.
 - **Flat shape** (`task.batch` off): `{ agent, ...item }` — exactly one spawn per call. Shared background goes into a `local://` file (e.g. `local://ctx.md`) that each assignment references; subagents share the parent's `local://` root.
 
 | Field | Type | Required | Description |
@@ -37,6 +37,7 @@ The wire schema is shape-swapped by `task.batch` (default on). One unit of work
 | `context` | `string` | Yes (batch) | Shared background prepended to every spawn of the call via the subagent system prompt. Rejected when `task.batch` is off. |
 | `tasks` | `array` | Yes (batch) | One task item per subagent. Provided ids must be unique within the call (case-insensitive). Rejected when `task.batch` is off. |
 | `id` | `string` | No | Stable agent id, schema max length 48. Defaults to a generated AdjectiveNoun name. Uniquified per session by `AgentOutputManager`. Item field in batch shape, top-level in flat shape. |
+| `role` | `string` | No | Specialist role/expertise for this spawn, schema max length 256. Shapes the subagent system-prompt identity and roster display name; generic `task` / `quick_task` spawns without roles get an advisory nudge. Item field in batch shape, top-level in flat shape. |
 | `description` | `string` | No | UI label only; the subagent never sees it. Item field in batch shape, top-level in flat shape. |
 | `assignment` | `string` | Yes | The work — complete, self-contained instructions. Empty-after-trim is rejected. Item field in batch shape, top-level in flat shape. |
 | `isolated` | `boolean` | No | Run in an isolated workspace and return patches. Exists only when `task.isolation.mode` is not `none`; per item in batch shape, top-level in flat shape. Isolated agents are torn down at completion — not revivable. |