selftune 0.2.22 → 0.2.24

package/packages/dashboard-core/src/screens/skills/SkillsLibraryScreen.tsx

@@ -0,0 +1,162 @@
+"use client";
+
+import type { ReactNode } from "react";
+import { useMemo, useState } from "react";
+
+import type { DerivedSkill, FilterTab } from "@selftune/ui/components";
+import {
+  LibraryHealthCard,
+  PendingProposalsCard,
+  SkillCardItem,
+  SkillFilterTabs,
+  SkillGridEmpty,
+  SkillHeroCard,
+  SkillHeroEmpty,
+  SkillsLibraryError,
+  SkillsLibrarySkeleton,
+} from "@selftune/ui/components";
+
+export interface SkillsLibraryHero {
+  skillName: string;
+  skillScope?: string | null;
+  platforms?: string[];
+  passRate: number | null;
+  totalChecks: number;
+  uniqueSessions: number;
+  status: DerivedSkill["status"];
+  latestEvolutionTimestamp?: string | null;
+}
+
+export interface SkillsLibraryPendingProposal {
+  id: string;
+  skillName: string | null;
+  action: string;
+}
+
+export interface SkillsLibraryScreenProps {
+  skills: DerivedSkill[];
+  heroSkill?: SkillsLibraryHero | null;
+  aggregatePassRate: number | null;
+  gradedCount: number;
+  pendingProposals: SkillsLibraryPendingProposal[];
+  isLoading: boolean;
+  error?: string | null;
+  onRetry(): void;
+  renderHeroActions(skillName: string): ReactNode;
+  renderCardActions(skillName: string): ReactNode;
+}
+
+export function SkillsLibraryScreen({
+  skills,
+  heroSkill,
+  aggregatePassRate,
+  gradedCount,
+  pendingProposals,
+  isLoading,
+  error,
+  onRetry,
+  renderHeroActions,
+  renderCardActions,
+}: SkillsLibraryScreenProps) {
+  const [filter, setFilter] = useState<FilterTab>("ALL");
+  const [sortDesc, setSortDesc] = useState(true);
+
+  const filteredSkills = useMemo(() => {
+    let result = skills;
+    if (filter !== "ALL") {
+      result = result.filter((skill) => skill.status === filter);
+    }
+    if (!sortDesc) {
+      return result;
+    }
+    return result.reduceRight<DerivedSkill[]>((acc, skill) => {
+      acc.push(skill);
+      return acc;
+    }, []);
+  }, [filter, skills, sortDesc]);
+
+  const counts = useMemo<Record<FilterTab, number>>(() => {
+    const nextCounts: Record<FilterTab, number> = {
+      ALL: skills.length,
+      HEALTHY: 0,
+      WARNING: 0,
+      CRITICAL: 0,
+      UNGRADED: 0,
+    };
+
+    for (const skill of skills) {
+      if (skill.status in nextCounts) {
+        nextCounts[skill.status as Exclude<FilterTab, "ALL">]++;
+      }
+    }
+
+    return nextCounts;
+  }, [skills]);
+
+  if (isLoading) {
+    return <SkillsLibrarySkeleton />;
+  }
+
+  if (error) {
+    return <SkillsLibraryError message={error} onRetry={onRetry} />;
+  }
+
+  return (
+    <div
+      data-parity-root="skills-library"
+      className="@container/main flex flex-1 animate-in fade-in flex-col gap-8 px-4 py-8 duration-500 lg:px-6"
+    >
+      <div>
+        <h1 className="font-headline text-4xl font-bold tracking-tight text-foreground">
+          Skills Library
+        </h1>
+        <p className="mt-2 max-w-2xl text-sm text-muted-foreground">
+          Monitor and manage your evolving skill definitions across all scopes.
+        </p>
+      </div>
+
+      <div className="grid grid-cols-12 gap-6">
+        {heroSkill ? (
+          <SkillHeroCard
+            skillName={heroSkill.skillName}
+            skillScope={heroSkill.skillScope ?? null}
+            platforms={heroSkill.platforms}
+            passRate={heroSkill.passRate}
+            totalChecks={heroSkill.totalChecks}
+            uniqueSessions={heroSkill.uniqueSessions}
+            status={heroSkill.status}
+            latestEvolutionTimestamp={heroSkill.latestEvolutionTimestamp ?? null}
+            renderActions={renderHeroActions}
+          />
+        ) : (
+          <SkillHeroEmpty />
+        )}
+
+        <div className="col-span-12 flex flex-col gap-6 lg:col-span-4">
+          <LibraryHealthCard aggregatePassRate={aggregatePassRate} gradedCount={gradedCount} />
+          <PendingProposalsCard proposals={pendingProposals} />
+        </div>
+      </div>
+
+      <div className="space-y-6">
+        <SkillFilterTabs
+          filter={filter}
+          onFilterChange={setFilter}
+          counts={counts}
+          sortDesc={sortDesc}
+          onSortToggle={() => setSortDesc((value) => !value)}
+        />
+
+        {filteredSkills.length > 0 ? (
+          <div className="grid grid-cols-1 gap-6 md:grid-cols-2 xl:grid-cols-3">
+            {filteredSkills.map((skill) => (
+              <SkillCardItem key={skill.name} skill={skill} renderActions={renderCardActions} />
+            ))}
+          </div>
+        ) : (
+          <SkillGridEmpty />
+        )}
+      </div>
+    </div>
+  );
+}

package/packages/dashboard-core/src/screens/skills/index.ts

@@ -0,0 +1,6 @@
+export {
+  SkillsLibraryScreen,
+  type SkillsLibraryHero,
+  type SkillsLibraryPendingProposal,
+  type SkillsLibraryScreenProps,
+} from "./SkillsLibraryScreen";

package/packages/telemetry-contract/fixtures/complete-push.ts

@@ -1,4 +1,4 @@
-import type { PushPayloadV2 } from "../src/schemas.js";
+import type { PushPayloadV2 } from "../src/types.js";
 
 /**
  * A valid PushPayloadV2 with at least one of every record type.

package/packages/telemetry-contract/fixtures/evidence-only-push.ts

@@ -1,4 +1,4 @@
-import type { PushPayloadV2 } from "../src/schemas.js";
+import type { PushPayloadV2 } from "../src/types.js";
 
 /**
  * A valid PushPayloadV2 with only evolution_evidence entries and
@@ -7,7 +7,7 @@ import type { PushPayloadV2 } from "../src/schemas.js";
 export const evidenceOnlyPush: PushPayloadV2 = {
   schema_version: "2.0",
   client_version: "0.9.0",
-  push_id: "d4e5f6a7-b8c9-0123-defa-234567890123",
+  push_id: "d4e5f6a7-b8c9-8123-9efa-234567890123",
   normalizer_version: "0.2.1",
   canonical: {
     sessions: [],

package/packages/telemetry-contract/fixtures/golden.test.ts

@@ -1,7 +1,6 @@
 import { describe, expect, test } from "bun:test";
 import { readFileSync } from "node:fs";
 import { join } from "node:path";
-
 import { CANONICAL_SCHEMA_VERSION } from "../src/types.js";
 import { isCanonicalRecord } from "../src/validators.js";
 

package/packages/telemetry-contract/fixtures/partial-push-no-sessions.ts

@@ -1,4 +1,4 @@
-import type { PushPayloadV2 } from "../src/schemas.js";
+import type { PushPayloadV2 } from "../src/types.js";
 
 /**
  * A valid PushPayloadV2 with zero sessions but non-empty evolution_evidence.

package/packages/telemetry-contract/fixtures/partial-push-unresolved-parents.ts

@@ -1,4 +1,4 @@
-import type { PushPayloadV2 } from "../src/schemas.js";
+import type { PushPayloadV2 } from "../src/types.js";
 
 /**
  * A valid PushPayloadV2 with invocations and prompts that reference a
@@ -10,7 +10,7 @@ import type { PushPayloadV2 } from "../src/schemas.js";
 export const partialPushUnresolvedParents: PushPayloadV2 = {
   schema_version: "2.0",
   client_version: "0.9.0",
-  push_id: "c3d4e5f6-a7b8-9012-cdef-123456789012",
+  push_id: "c3d4e5f6-a7b8-8012-8def-123456789012",
   normalizer_version: "0.2.1",
   canonical: {
     sessions: [],

package/packages/telemetry-contract/package.json

@@ -13,9 +13,9 @@
   "type": "module",
   "exports": {
     ".": "./index.ts",
-    "./schemas": "./src/schemas.ts",
     "./types": "./src/types.ts",
     "./validators": "./src/validators.ts",
+    "./schemas": "./src/schemas.ts",
     "./fixtures": "./fixtures/index.ts"
   },
   "dependencies": {

package/packages/telemetry-contract/src/index.ts

@@ -1,2 +1,3 @@
+export * from "./schemas.js";
 export * from "./types.js";
 export * from "./validators.js";

package/packages/telemetry-contract/src/schemas.ts

@@ -1,5 +1,13 @@
-import { z } from "zod";
+/**
+ * Zod validation schemas for all canonical telemetry record types
+ * and the PushPayloadV2 envelope.
+ *
+ * This is the single source of truth -- cloud consumers should import
+ * from @selftune/telemetry-contract/schemas instead of maintaining
+ * their own copies.
+ */
 
+import { z } from "zod";
 import {
   CANONICAL_CAPTURE_MODES,
   CANONICAL_COMPLETION_STATUSES,
@@ -11,6 +19,8 @@ import {
   CANONICAL_SOURCE_SESSION_KINDS,
 } from "./types.js";
 
+// ---------- Shared enum schemas ----------
+
 export const canonicalPlatformSchema = z.enum(CANONICAL_PLATFORMS);
 export const captureModeSchema = z.enum(CANONICAL_CAPTURE_MODES);
 export const sourceSessionKindSchema = z.enum(CANONICAL_SOURCE_SESSION_KINDS);
@@ -19,6 +29,8 @@ export const invocationModeSchema = z.enum(CANONICAL_INVOCATION_MODES);
 export const completionStatusSchema = z.enum(CANONICAL_COMPLETION_STATUSES);
 export const recordKindSchema = z.enum(CANONICAL_RECORD_KINDS);
 
+// ---------- Shared structural schemas ----------
+
 export const rawSourceRefSchema = z.object({
   path: z.string().optional(),
   line: z.number().int().nonnegative().optional(),
@@ -42,6 +54,8 @@ export const canonicalSessionRecordBaseSchema = canonicalRecordBaseSchema.extend
   session_id: z.string().min(1),
 });
 
+// ---------- Canonical record schemas ----------
+
 export const CanonicalSessionRecordSchema = canonicalSessionRecordBaseSchema.extend({
   record_kind: z.literal("session"),
   external_session_id: z.string().optional(),
@@ -126,10 +140,13 @@ export const CanonicalEvolutionEvidenceRecordSchema = z.object({
   evidence_id: z.string().min(1),
   skill_name: z.string().min(1),
   proposal_id: z.string().optional(),
+  timestamp: z.string().datetime().optional(),
+  skill_path: z.string().optional(),
   target: z.string().min(1),
   stage: z.string().min(1),
   rationale: z.string().optional(),
   confidence: z.number().min(0).max(1).optional(),
+  details: z.string().optional(),
   original_text: z.string().optional(),
   proposed_text: z.string().optional(),
   eval_set_json: z.unknown().optional(),
@@ -137,6 +154,39 @@ export const CanonicalEvolutionEvidenceRecordSchema = z.object({
   raw_source_ref: rawSourceRefSchema.optional(),
 });
 
+export const CanonicalGradingResultRecordSchema = z.object({
+  grading_id: z.string().min(1),
+  session_id: z.string().min(1),
+  skill_name: z.string().min(1),
+  transcript_path: z.string().nullable().optional(),
+  graded_at: z.string().min(1),
+  pass_rate: z.number().min(0).max(1).nullable().optional(),
+  mean_score: z.number().min(0).max(1).nullable().optional(),
+  score_std_dev: z.number().nullable().optional(),
+  passed_count: z.number().int().nonnegative().nullable().optional(),
+  failed_count: z.number().int().nonnegative().nullable().optional(),
+  total_count: z.number().int().nonnegative().nullable().optional(),
+  expectations_json: z.string().nullable().optional(),
+  claims_json: z.string().nullable().optional(),
+  eval_feedback_json: z.string().nullable().optional(),
+  failure_feedback_json: z.string().nullable().optional(),
+  execution_metrics_json: z.string().nullable().optional(),
+});
+
+export const CanonicalImprovementSignalRecordSchema = z.object({
+  signal_id: z.string().min(1),
+  timestamp: z.string().min(1),
+  session_id: z.string().min(1),
+  query: z.string().min(1),
+  signal_type: z.string().min(1),
+  mentioned_skill: z.string().nullable().optional(),
+  consumed: z.boolean(),
+  consumed_at: z.string().nullable().optional(),
+  consumed_by_run: z.string().nullable().optional(),
+});
+
+// ---------- Orchestrate run schemas ----------
+
 export const OrchestrateRunSkillActionSchema = z.object({
   skill: z.string().min(1),
   action: z.enum(["evolve", "watch", "skip"]),
@@ -163,12 +213,12 @@ export const PushOrchestrateRunRecordSchema = z.object({
   skill_actions: z.array(OrchestrateRunSkillActionSchema),
 });
 
+// ---------- Push V2 envelope ----------
+
 export const PushPayloadV2Schema = z.object({
   schema_version: z.literal("2.0"),
   client_version: z.string().min(1),
-  // Queue-generated push IDs are typically UUIDs, but the wire contract only
-  // requires a stable non-empty idempotency key.
-  push_id: z.string().min(1),
+  push_id: z.string().uuid(),
   normalizer_version: z.string().min(1),
   canonical: z.object({
     sessions: z.array(CanonicalSessionRecordSchema).min(0),
@@ -178,10 +228,14 @@ export const PushPayloadV2Schema = z.object({
     normalization_runs: z.array(CanonicalNormalizationRunRecordSchema).min(0),
     evolution_evidence: z.array(CanonicalEvolutionEvidenceRecordSchema).optional(),
     orchestrate_runs: z.array(PushOrchestrateRunRecordSchema).optional(),
+    grading_results: z.array(CanonicalGradingResultRecordSchema).optional(),
+    improvement_signals: z.array(CanonicalImprovementSignalRecordSchema).optional(),
   }),
 });
 
-export type PushPayloadV2 = z.infer<typeof PushPayloadV2Schema>;
+// ---------- Inferred types from Zod schemas ----------
+
+export type ZodPushPayloadV2 = z.infer<typeof PushPayloadV2Schema>;
 export type ZodCanonicalSessionRecord = z.infer<typeof CanonicalSessionRecordSchema>;
 export type ZodCanonicalPromptRecord = z.infer<typeof CanonicalPromptRecordSchema>;
 export type ZodCanonicalSkillInvocationRecord = z.infer<
@@ -194,4 +248,8 @@ export type ZodCanonicalNormalizationRunRecord = z.infer<
 export type ZodCanonicalEvolutionEvidenceRecord = z.infer<
   typeof CanonicalEvolutionEvidenceRecordSchema
 >;
+export type ZodCanonicalGradingResultRecord = z.infer<typeof CanonicalGradingResultRecordSchema>;
+export type ZodCanonicalImprovementSignalRecord = z.infer<
+  typeof CanonicalImprovementSignalRecordSchema
+>;
 export type ZodPushOrchestrateRunRecord = z.infer<typeof PushOrchestrateRunRecordSchema>;

package/packages/telemetry-contract/src/types.ts

@@ -1,7 +1,7 @@
 export const CANONICAL_SCHEMA_VERSION = "2.0" as const;
 export type CanonicalSchemaVersion = typeof CANONICAL_SCHEMA_VERSION;
 
-export const CANONICAL_PLATFORMS = ["claude_code", "codex", "opencode", "openclaw"] as const;
+export const CANONICAL_PLATFORMS = ["claude_code", "codex", "opencode", "openclaw", "pi"] as const;
 export type CanonicalPlatform = (typeof CANONICAL_PLATFORMS)[number];
 
 export const CANONICAL_CAPTURE_MODES = [
@@ -54,7 +54,7 @@ export const CANONICAL_RECORD_KINDS = [
 ] as const;
 export type CanonicalRecordKind = (typeof CANONICAL_RECORD_KINDS)[number];
 
-export interface CanonicalRawSourceRef {
+export interface CanonicalRawSourceRef extends Record<string, unknown> {
   path?: string;
   line?: number;
   event_type?: string;
@@ -62,7 +62,7 @@ export interface CanonicalRawSourceRef {
   metadata?: Record<string, unknown>;
 }
 
-export interface CanonicalRecordBase {
+export interface CanonicalRecordBase extends Record<string, unknown> {
   record_kind: CanonicalRecordKind;
   schema_version: CanonicalSchemaVersion;
   normalizer_version: string;
@@ -146,15 +146,14 @@ export interface CanonicalExecutionFactRecord extends CanonicalSessionRecordBase
   cached_input_tokens?: number;
   reasoning_output_tokens?: number;
   cost_usd?: number;
-  duration_ms?: number;
   files_changed?: number;
   lines_added?: number;
   lines_removed?: number;
   lines_modified?: number;
-  /** Count of output-producing tool calls (Write, Edit, WebFetch, WebSearch, Skill, Agent). */
   artifact_count?: number;
-  /** Inferred session type based on tool distribution. */
-  session_type?: "dev" | "research" | "content" | "mixed";
+  session_type?: string;
+  agent_summary?: string;
+  duration_ms?: number;
   completion_status?: CanonicalCompletionStatus;
   end_reason?: string;
 }
@@ -168,9 +167,100 @@ export interface CanonicalNormalizationRunRecord extends CanonicalRecordBase {
   repair_applied: boolean;
 }
 
+export interface CanonicalEvolutionEvidenceRecord {
+  evidence_id?: string;
+  timestamp?: string;
+  proposal_id?: string;
+  skill_name: string;
+  skill_path?: string;
+  target: string;
+  stage: string;
+  rationale?: string;
+  confidence?: number;
+  details?: string;
+  original_text?: string;
+  proposed_text?: string;
+  eval_set_json?: unknown;
+  validation_json?: unknown;
+  raw_source_ref?: CanonicalRawSourceRef;
+}
+
+export interface CanonicalGradingResultRecord {
+  grading_id: string;
+  session_id: string;
+  skill_name: string;
+  transcript_path?: string | null;
+  graded_at: string;
+  pass_rate?: number | null;
+  mean_score?: number | null;
+  score_std_dev?: number | null;
+  passed_count?: number | null;
+  failed_count?: number | null;
+  total_count?: number | null;
+  expectations_json?: string | null;
+  claims_json?: string | null;
+  eval_feedback_json?: string | null;
+  failure_feedback_json?: string | null;
+  execution_metrics_json?: string | null;
+}
+
+export interface CanonicalImprovementSignalRecord {
+  signal_id: string;
+  timestamp: string;
+  session_id: string;
+  query: string;
+  signal_type: string;
+  mentioned_skill?: string | null;
+  consumed: boolean;
+  consumed_at?: string | null;
+  consumed_by_run?: string | null;
+}
+
 export type CanonicalRecord =
   | CanonicalSessionRecord
   | CanonicalPromptRecord
   | CanonicalSkillInvocationRecord
   | CanonicalExecutionFactRecord
   | CanonicalNormalizationRunRecord;
+
+export interface PushOrchestrateRunRecord {
+  run_id: string;
+  timestamp: string;
+  elapsed_ms: number;
+  dry_run: boolean;
+  approval_mode: "auto" | "review";
+  total_skills: number;
+  evaluated: number;
+  evolved: number;
+  deployed: number;
+  watched: number;
+  skipped: number;
+  skill_actions: Array<{
+    skill: string;
+    action: "evolve" | "watch" | "skip";
+    reason: string;
+    deployed?: boolean;
+    rolledBack?: boolean;
+    alert?: string | null;
+    elapsed_ms?: number;
+    llm_calls?: number;
+  }>;
+}
+
+export interface PushPayloadV2 {
+  schema_version: CanonicalSchemaVersion;
+  client_version: string;
+  push_id: string;
+  normalizer_version: string;
+  canonical: {
+    sessions: CanonicalSessionRecord[];
+    prompts: CanonicalPromptRecord[];
+    skill_invocations: CanonicalSkillInvocationRecord[];
+    execution_facts: CanonicalExecutionFactRecord[];
+    normalization_runs: CanonicalNormalizationRunRecord[];
+    evolution_evidence?: CanonicalEvolutionEvidenceRecord[];
+    orchestrate_runs?: PushOrchestrateRunRecord[];
+    grading_results?: CanonicalGradingResultRecord[];
+    improvement_signals?: CanonicalImprovementSignalRecord[];
+  };
+}

package/packages/telemetry-contract/tests/compatibility.test.ts

@@ -1,5 +1,4 @@
 import { describe, expect, test } from "bun:test";
-
 import { completePush } from "../fixtures/complete-push.js";
 import { evidenceOnlyPush } from "../fixtures/evidence-only-push.js";
 import { partialPushNoSessions } from "../fixtures/partial-push-no-sessions.js";

package/packages/ui/AGENTS.md

@@ -0,0 +1,16 @@
+# @selftune/ui
+
+Shared React component library used by both the cloud dashboard and the local OSS dashboard. Canonical copy lives here; synced to `oss/selftune/packages/ui` via `scripts/sync-embedded-shared.sh`.
+
+| Directory         | Contents                                                                                                                                                                                                                                    |
+| ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| `src/primitives/` | Base UI: Badge, Button, Card, Checkbox, Collapsible, DropdownMenu, Label, Select, Table, Tabs, Tooltip                                                                                                                                      |
+| `src/components/` | Shared components: SkillHealthGrid, EvolutionTimeline, EvidenceViewer, ActivityTimeline, OrchestrateRunsPanel, SectionCards, InfoTip, SkillReportPanels, SkillReportGuide, InvocationsPanel, SkillsLibrary, AnalyticsCharts, OverviewPanels |
+| `src/types.ts`    | Shared types: SkillCard, SkillHealthStatus, EvalSnapshot, EvolutionEntry, TrustState, TrustFields, ExampleRow, AutonomyStatus, TrustWatchlistEntry, AttentionItem, AutonomousDecision                                                       |
+| `src/lib/`        | Utilities: format (formatRate, timeAgo), constants (STATUS_CONFIG), utils (deriveStatus, sortByPassRateAndChecks)                                                                                                                           |
+
+**Exports:** `./primitives`, `./components`, `./types`, `./lib`
+
+**Dependencies:** react, @base-ui/react, lucide-react, clsx, tailwind-merge
+
+**Important:** Do NOT edit `oss/selftune/packages/ui/` directly. Edit here and run `scripts/sync-embedded-shared.sh`.

package/packages/ui/README.md

@@ -57,7 +57,7 @@ Presentational components for selftune dashboard views. No data fetching, no rou
 | ---------------------- | -------------------------------------------------------------------------------------------------------------------------------- |
 | `SkillHealthGrid`      | Sortable/filterable data table with drag-and-drop, pagination, and view tabs. Accepts `renderSkillName` prop for custom routing. |
 | `EvolutionTimeline`    | Proposal lifecycle timeline grouped by proposal ID, with pass rate deltas.                                                       |
-| `ActivityPanel`        | Tabbed activity feed (undeployed proposals, timeline events, unmatched queries).                                                 |
+| `ActivityPanel`        | Tabbed activity feed (pending proposals, timeline events, unmatched queries).                                                    |
 | `EvidenceViewer`       | Full evidence trail for a proposal — side-by-side diffs, validation results, iteration rounds.                                   |
 | `SectionCards`         | Dashboard metric stat cards (skills count, pass rate, unmatched, sessions, etc.).                                                |
 | `OrchestrateRunsPanel` | Collapsible orchestrate run reports with per-skill action details.                                                               |

package/packages/ui/package.json

@@ -38,7 +38,7 @@
     "react": "^19.0.0",
     "react-dom": "^19.0.0",
     "react-markdown": "^10.0.0",
-    "recharts": "^2.0.0"
+    "recharts": "^3.0.0"
   },
   "peerDependenciesMeta": {
     "@dnd-kit/core": {