@caupulican/pi-adaptative 0.80.96 → 0.80.98

package/CHANGELOG.md

@@ -1,7 +1,42 @@
+## [0.80.98] - 2026-07-02
+
+### Added
+
+- Added the executor lane (G16): configure `modelRouter.executorModel` (or assign it in one step
+  from /fitness -> "Toolkit executor") and command-shaped prompts that score a deterministic
+  Level-0 EXACT hit on the toolkit registry route the whole turn to the local executor — with the
+  judge skipped (nothing to judge), the tool surface capability-filtered to the executor's own
+  class, and `run_toolkit_script` exempted from the cheap-tier mutation escalation ONLY on these
+  routes (its own danger gate still applies; any other mutating tool still escalates). Ambiguous
+  requests never route here — they stay with the main model and the reflex brain.
+- Added workspace research sources (G1): the autonomous research lane now grounds itself in the
+  repo — a bounded ripgrep collector derives terms from the goal text and feeds pointer-first
+  sources (repo-relative path, line, <=200-char excerpt; never file bodies) into the research
+  prompt and the evidence bundle, with the synthesis anchors always preserved. Best-effort: no
+  ripgrep or no matches reproduces the previous behavior exactly.
+- Added real evidence strength to the learning gate (G6): a bounded persistent observation store
+  counts how often the same lesson (layer + normalized summary) is re-observed across passes and
+  sessions, so `minObservations` now gates on truth — the first observation proposals, repeated
+  ones can auto-apply.
+- Added autonomy telemetry emission (G3): route decisions, research/worker lane outcomes, and
+  learning decisions now emit redacted, bounded events (ids/codes/numbers only — never prompt or
+  memory text) as `autonomy-telemetry` session entries.
+- Added execution-time capability-envelope path enforcement and worker request persistence (G2
+  prerequisites): a tool wrapped in an envelope scope structurally refuses out-of-scope paths at
+  the moment it runs (deny-wins, escape-proof), and every worker result persists its originating
+  request (instructions, route, envelope) for audit.
+- Added digest-served telemetry: the /context curation line now reports how many brain digests
+  were actually rendered into GC stubs on real turns — the pays-for-itself number.
+
+## [0.80.97] - 2026-07-02
+
 ## [0.80.96] - 2026-07-02
 
 ### Added
 
+- Added Settings submenus for Learning Policy (enabled, auto-apply, confidence threshold, min
+  observations; layer allow-list stays JSON-only) and Model Capability (auto/off/forced class),
+  closing the last two settings surfaces that were configurable only by hand-editing JSON (G5).
 - Added routed-turn capability tool filtering (G4): when the model router swaps a turn to a
   cheaper model, the tool surface now follows the ROUTED model's capability class for that turn
   (an 8k local model no longer inherits — and pays schema tokens for — background-autonomy tools

package/dist/core/agent-session.d.ts

@@ -1,6 +1,6 @@
 import type { Agent, AgentEvent, AgentMessage, AgentState, AgentTool, ThinkingLevel } from "@caupulican/pi-agent-core";
 import type { CacheRetention, ImageContent, Message, Model, StopReason, TextContent, Usage } from "@caupulican/pi-ai";
-import type { CapabilityEnvelope, EvidenceBundle, LearningDecision, WorkerResult } from "./autonomy/contracts.ts";
+import type { CapabilityEnvelope, EvidenceBundle, LearningDecision, WorkerRequest, WorkerResult } from "./autonomy/contracts.ts";
 import { type LaneRecord } from "./autonomy/lane-tracker.ts";
 import type { AutonomyDiagnosticSnapshot, AutonomyStatusSnapshot } from "./autonomy/status.ts";
 import { type BashResult } from "./bash-executor.ts";
@@ -32,6 +32,7 @@ import { type StoredFitnessReport } from "./models/fitness-store.ts";
 import { type PromptTemplate } from "./prompt-templates.ts";
 import { type ModelFitnessReport } from "./research/model-fitness.ts";
 import { type ResearchRunResult } from "./research/research-runner.ts";
+import { collectWorkspaceSources } from "./research/workspace-collector.ts";
 import type { ResourceLoader } from "./resource-loader.ts";
 import type { BranchSummaryEntry, SessionManager } from "./session-manager.ts";
 import { type ResourceProfileFilterSettings, type SettingsManager } from "./settings-manager.ts";
@@ -141,6 +142,12 @@ export interface AgentSessionConfig {
     };
     /** Session start event metadata emitted when extensions bind to this runtime. */
     sessionStartEvent?: SessionStartEvent;
+    /**
+     * Pointer-first workspace source collector for the autonomous research lane. Injected in unit
+     * tests so they don't spawn a real ripgrep child (which would escape fake timers); production
+     * defaults to the real, best-effort collector.
+     */
+    collectWorkspaceSources?: typeof collectWorkspaceSources;
 }
 export interface ExtensionBindings {
     uiContext?: ExtensionUIContext;
@@ -362,6 +369,7 @@ export declare class AgentSession {
     private _baseToolDefinitions;
     private _cwd;
     private _agentDir;
+    private _collectWorkspaceSources;
     private _extensionRunnerRef?;
     private _initialActiveToolNames?;
     private _allowedToolNames?;
@@ -781,6 +789,7 @@ export declare class AgentSession {
     private _rebuildSystemPrompt;
     private _runAgentPrompt;
     private _isModelAvailableAndAuthed;
+    private _resolveExecutorRoute;
     private _resolveModelRouterTurnRoute;
     private _resolveModelRouterModelForIntent;
     private _resolveConfiguredTierModel;
@@ -1182,7 +1191,15 @@ export declare class AgentSession {
     getEvidenceBundleSnapshots(): EvidenceBundle[];
     /** Live lane records tracked by this process (running and terminal). */
     getLaneRecords(): LaneRecord[];
-    saveWorkerResultSnapshot(result: WorkerResult): string;
+    /**
+     * G3: bounded autonomy-telemetry sink. Passes the whole event through {@link redactTelemetryValue}
+     * (the taxonomy's redaction contract) before storing it, so a secret that leaked into a payload
+     * field never lands in the session log. Observe-only: a failure here can never surface into the
+     * turn it is measuring, so the whole body is swallowed. Payloads MUST stay small (ids, codes,
+     * numbers) — never prompt/summary text; callers own that discipline.
+     */
+    private _emitAutonomyTelemetry;
+    saveWorkerResultSnapshot(result: WorkerResult, request?: WorkerRequest): string;
     getWorkerResultSnapshots(): WorkerResult[];
     saveLearningDecisionSnapshot(decision: LearningDecision): string;
     getLearningDecisionSnapshots(): LearningDecision[];