sortie-dogs 0.4.7 → 0.5.0

package/README.md

@@ -11,16 +11,16 @@
 
 ![Sortie-dogs coordinating a bounded implementation workflow](https://raw.githubusercontent.com/zufall-upon/Sortie-dogs/main/docs/assets/sortie-workflow.gif)
 
-Sortie-dogs is an opt-in OpenCode orchestration plugin. It turns a task into a
-scoped plan, parallel investigation, bounded independent implementation, canonical
-validation, and evidence-backed completion—while preserving standard OpenCode
-agents and settings.
+Sortie-dogs is an opt-in OpenCode orchestration plugin. It turns a task into a
+scoped plan, optional evidence gathering, one bounded implementation unit, canonical
+validation, and evidence-backed completion—while preserving standard OpenCode
+agents and settings.
 
 Requirements: Node.js 22.6 or newer, npm, and OpenCode.
 
 Guides: [日本語](docs/guide-ja.md) · [简体中文](docs/guide-zh-CN.md) · [CLI testing](docs/cli-testing.md)
 
-Release: [v0.4.7](https://github.com/zufall-upon/Sortie-dogs/releases/tag/v0.4.7)
+Release: [v0.5.0](https://github.com/zufall-upon/Sortie-dogs/releases/tag/v0.5.0)
 
 ## Quick start
 
@@ -62,7 +62,7 @@ to the `plugin` array of the OpenCode configuration the agents run under —
 Restart OpenCode afterwards. A `plugin` entry must name the package, not a
 subpath: `sortie-dogs/plugin` is an import specifier, not a plugin specifier.
 
-`dog-coordinator` defaults to `openai/gpt-5.6-terra` with the `medium` variant; `dog-scout` defaults to
+`dog-coordinator` defaults to `openai/gpt-5.6-luna` with the `max` variant; `dog-scout` defaults to
 `openai/gpt-5.6-luna`. To pin either role to another model, save this
 as `.opencode/sortie-dogs.json`:
 
@@ -203,12 +203,14 @@ global file for durable global settings.
 
 - **Focused when invited, invisible otherwise.** Activate it with `/sortie` or
   select `dog-coordinator`; ordinary OpenCode sessions remain unchanged.
-- **Parallel context without uncontrolled fan-out.** Every worker handoff uses
-  exactly three bounded scouts before implementation begins.
+- **Evidence gathering without arbitrary dispatch ceilings.** A bounded scout resolves
+  each concrete manifest, validation, or ownership gap whenever it appears; unchanged
+  requests are not repeated.
 - **Writes stay inside the assignment.** Exact source or operation manifests
   gate edits and handoffs.
-- **Safe implementation fan-out.** Independent units may use two or three workers
-  with non-overlapping write manifests; dependent or same-path work stays on one worker.
+- **Autonomous sequential implementation.** A normal request may use as many sequential
+  workers as its accepted scope requires. Concurrent fan-out still needs the explicit
+  parallel contract.
 - **Runtime overlap protection.** Active equal or ancestor write scopes are rejected
   before mutation, and full validation waits until every parallel unit joins.
 - **Evidence before completion.** Canonical validation, risk-based review, and
@@ -223,8 +225,8 @@ An illustrative low-risk run stays bounded and reports its gates:
 ```text
 You: /sortie Add the requested behavior
 dog-coordinator: manifest confirmed
-dog-scout ×3: investigation complete
-dog-worker: implementation complete
+dog-scout: skipped — no concrete evidence gap
+dog-worker: implementation complete
 validation: npm test — PASS
 review: skipped — low risk
 dog-coordinator: completion evidence accepted
@@ -234,8 +236,8 @@ dog-coordinator: completion evidence accepted
 
 1. **Brief and plan** — `dog-coordinator` turns the request into acceptance
    criteria, a write manifest, and validation requirements.
-2. **Exactly three scouts** — bounded, read-only investigation collects
-   complementary evidence without expanding the write scope.
+2. **Optional scout** — one bounded, read-only investigation runs only for a
+   concrete pre-worker evidence gap.
 3. **Dedicated worker** — the dedicated worker implements only the approved
    manifest and also owns scoped remediation or blocker resolution.
 4. **Canonical validation** — the declared test or build command must produce
@@ -288,13 +290,13 @@ untouched.
 ## Model routing
 
 Default routes split work by required capability and repeated-context cost.
-Sortie-dogs keeps retrieval on Luna, coordinator state and routing on Terra,
-and independent review on Sol unless the host declares another target.
+Sortie-dogs keeps retrieval and coordinator routing on Luna, and independent
+review on Sol unless the host declares another target.
 
-`dog-coordinator` defaults to `openai/gpt-5.6-terra` with the `medium` variant. Coordinator work repeatedly
-processes the root context but is primarily state management and routing, so
-Terra is the cost-conscious middle tier between Luna and Sol. Project or global
-`modelRouting` can override this default. If the host proves Terra unavailable,
+`dog-coordinator` defaults to `openai/gpt-5.6-luna` with the `max` variant. Coordinator work repeatedly
+processes the root context, so Luna Max keeps deliberation high without paying
+Terra rates for every cached turn. Project or global `modelRouting` can override
+this default. If the host proves Luna unavailable,
 the existing availability policy uses a configured free-tier fallback when present
 and otherwise preserves the session model.
 
@@ -332,7 +334,7 @@ actually serve.
 {
   "modelRouting": {
     "dog-coordinator": {
-      "preferred": { "model": "openai/gpt-5.6-terra", "variant": "medium" }
+      "preferred": { "model": "openai/gpt-5.6-luna", "variant": "max" }
     },
     "dog-scout": {
       "preferred": { "model": "openai/gpt-5.6-luna", "variant": "high" }
@@ -347,7 +349,6 @@ actually serve.
   },
   "modelCatalog": {
     "project": [
-      { "model": "openai/gpt-5.6-terra", "variants": ["medium"] },
       { "model": "openai/gpt-5.6-sol", "variants": ["medium", "xhigh"] },
       { "model": "openai/gpt-5.6-luna", "variants": ["max", "high"] },
       { "model": "anthropic/claude-opus-5" }

package/dist/asset-version.d.ts

@@ -2,5 +2,5 @@
  * Version of the installable runtime assets. Kept in its own module so the plugin can compare an
  * installed project marker without importing every asset body.
  */
-export declare const RUNTIME_ASSET_VERSION = "0.3.11-worker-resume-v1";
+export declare const RUNTIME_ASSET_VERSION = "0.3.32-terminal-outcome-v1";
 export type RuntimeAssetVersion = typeof RUNTIME_ASSET_VERSION;

package/dist/asset-version.js

@@ -2,4 +2,4 @@
  * Version of the installable runtime assets. Kept in its own module so the plugin can compare an
  * installed project marker without importing every asset body.
  */
-export const RUNTIME_ASSET_VERSION = "0.3.11-worker-resume-v1";
+export const RUNTIME_ASSET_VERSION = "0.3.32-terminal-outcome-v1";

package/dist/core/scope-lease-registry.d.ts

@@ -0,0 +1,53 @@
+import { type WorktreeScope } from "./worktree-scope.js";
+export type ScopeLeaseErrorCode = "scope-conflict" | "not-held" | "corrupt-state" | "invalid-request" | "lock-timeout" | "lock-lost";
+export declare class ScopeLeaseError extends Error {
+    readonly code: ScopeLeaseErrorCode;
+    constructor(code: ScopeLeaseErrorCode, message: string);
+}
+export type ScopeLeaseAcquireRequest = {
+    readonly scope: WorktreeScope;
+    readonly ownerId?: string;
+    readonly ttlMs?: number;
+};
+export type ScopeLeaseRegistryOptions = {
+    readonly ttlMs?: number;
+    readonly mutexRetries?: number;
+    readonly mutexRetryMs?: number;
+    readonly mutexStaleMs?: number;
+};
+export interface ScopeLease {
+    readonly id: string;
+    readonly ownerId: string;
+    readonly scope: WorktreeScope;
+    heartbeat(): Promise<void>;
+    assertHeld(): Promise<void>;
+    release(): Promise<void>;
+    abandon(): Promise<void>;
+    close(): void;
+}
+export declare class ScopeLeaseRegistry {
+    private readonly stateRoot;
+    private readonly statePath;
+    private readonly lockPath;
+    private readonly options;
+    constructor(stateRoot: string, options?: ScopeLeaseRegistryOptions);
+    static open(stateRoot: string, options?: ScopeLeaseRegistryOptions): ScopeLeaseRegistry;
+    acquire(request: ScopeLeaseAcquireRequest): Promise<ScopeLease>;
+    private validateRequest;
+    private makeLease;
+    private load;
+    private cleanupTemps;
+    private save;
+    private sameMutexSnapshot;
+    private sameDirectorySnapshot;
+    private mutexSnapshot;
+    private assertMutexHeld;
+    private quarantineSnapshot;
+    private sameQuarantineSnapshot;
+    private sameMovedQuarantineSnapshot;
+    private cleanupQuarantine;
+    private hasQuarantine;
+    private recoverStaleLock;
+    private removeOwnedMutex;
+    private withLock;
+}