@blackbelt-technology/pi-agent-dashboard 0.2.9 → 0.4.0

package/packages/extension/src/git-info.ts

@@ -1,7 +1,11 @@
 /**
  * Git info gathering — detects branch, remote URL, and PR number.
+ * Delegates to the shared git tool module so there's no inline execSync
+ * and every call benefits from the runner's safety defaults (windowsHide,
+ * timeout, tolerated exit codes).
+ * See change: platform-command-executor.
  */
-import { execSync } from "node:child_process";
+import * as git from "@blackbelt-technology/pi-dashboard-shared/platform/git.js";
 import { buildGitLinks, type GitLinks } from "./git-link-builder.js";
 
 export interface GitInfo {
@@ -11,39 +15,25 @@ export interface GitInfo {
   gitPrUrl?: string;
 }
 
-/** Run a shell command and return trimmed stdout, or undefined on failure. */
-function runGit(command: string, cwd: string): string | undefined {
-  try {
-    return execSync(command, { cwd, encoding: "utf-8", stdio: ["pipe", "pipe", "pipe"], timeout: 5000 }).trim();
-  } catch {
-    return undefined;
-  }
-}
-
 /** Detect the current git branch. Returns short SHA for detached HEAD. */
 export function detectBranch(cwd: string): string | undefined {
-  const ref = runGit("git rev-parse --abbrev-ref HEAD", cwd);
+  const ref = git.currentBranchOr({ cwd });
   if (!ref) return undefined;
   if (ref === "HEAD") {
     // Detached HEAD — return short commit SHA
-    return runGit("git rev-parse --short HEAD", cwd) ?? "HEAD";
+    return git.headShaOr({ cwd, short: true }) ?? "HEAD";
   }
   return ref;
 }
 
 /** Detect the remote origin URL. */
 export function detectRemoteUrl(cwd: string): string | undefined {
-  return runGit("git remote get-url origin", cwd);
+  return git.remoteUrlOr({ cwd });
 }
 
 /** Detect the PR number via gh CLI (best effort). */
 export function detectPrNumber(cwd: string): number | undefined {
-  const result = runGit("gh pr view --json number -q .number", cwd);
-  if (result) {
-    const num = parseInt(result, 10);
-    if (!isNaN(num)) return num;
-  }
-  return undefined;
+  return git.prNumberOr({ cwd });
 }
 
 /** Gather all git info for a directory. Returns undefined if not a git repo. */

package/packages/extension/src/pi-env.d.ts

@@ -30,6 +30,7 @@ declare module "@oh-my-pi/pi-coding-agent" {
     registerCommand(name: string, options: { description?: string; handler: (args: string, ctx: any) => Promise<void> }): void;
     registerTool(tool: any): void;
     registerProvider(name: string, config: any): void;
+    unregisterProvider(name: string): void;
     exec(command: string, args: string[], options?: { timeout?: number }): Promise<{ stdout: string; stderr: string; exitCode: number }>;
     events: EventBus;
   }

package/packages/extension/src/process-scanner.ts

@@ -12,8 +12,43 @@
  * This handles the reparenting problem: children get reparented to PID 1
  * when the bash wrapper exits, but we captured their PGIDs while alive.
  */
-import { spawnSync as defaultSpawnSync } from "node:child_process";
-import type { SpawnSyncReturns } from "node:child_process";
+import { spawnSync as defaultSpawnSync } from "@blackbelt-technology/pi-dashboard-shared/platform/exec.js";
+import type { SpawnSyncReturns } from "@blackbelt-technology/pi-dashboard-shared/platform/exec.js";
+import { getDefaultRegistry } from "@blackbelt-technology/pi-dashboard-shared/tool-registry/index.js";
+import { killPidWithGroup } from "@blackbelt-technology/pi-dashboard-shared/platform/process.js";
+
+/**
+ * Resolve a Windows system tool name (wmic / powershell / tasklist /
+ * taskkill) to its full `.exe` path via the global tool registry. If
+ * the registry lookup fails we fall back to the bare name and let
+ * `spawnSync` do PATHEXT resolution.
+ *
+ * Spawning the FULL path bypasses any cmd.exe / PATHEXT resolution
+ * layers, keeping `windowsHide: true` honored end-to-end. See change:
+ * consolidate-windows-spawn-and-platform-handlers.
+ *
+ * Uses `getDefaultRegistry` (not `peek*`) because the bridge extension
+ * runs inside pi's process and may be the FIRST caller to construct
+ * the registry in that process. Idempotent and cached per-process.
+ */
+const systemToolCache = new Map<string, string>();
+function resolveSystemTool(name: string): string {
+  const cached = systemToolCache.get(name);
+  if (cached) return cached;
+  try {
+    const reg = getDefaultRegistry();
+    if (reg.has(name)) {
+      const res = reg.resolve(name);
+      if (res.ok && res.path) {
+        systemToolCache.set(name, res.path);
+        return res.path;
+      }
+    }
+  } catch { /* registry unavailable in some test contexts */ }
+  // Cache the bare name so we only miss once per process.
+  systemToolCache.set(name, name);
+  return name;
+}
 
 export interface ChildProcessInfo {
   pid: number;
@@ -24,36 +59,12 @@ export interface ChildProcessInfo {
 
 /**
  * Parse ps ETIME format into milliseconds.
- * Formats: mm:ss, hh:mm:ss, dd-hh:mm:ss
+ * Re-exported from the shared platform primitive to keep the public API of
+ * this module stable while centralizing the pure helper.
+ * See change: consolidate-platform-handlers.
  */
-export function parseEtime(etime: string): number {
-  const trimmed = etime.trim();
-  if (!trimmed) return 0;
-
-  let days = 0;
-  let rest = trimmed;
-
-  const dashIdx = rest.indexOf("-");
-  if (dashIdx !== -1) {
-    days = parseInt(rest.slice(0, dashIdx), 10);
-    if (isNaN(days)) return 0;
-    rest = rest.slice(dashIdx + 1);
-  }
-
-  const parts = rest.split(":").map((p) => parseInt(p, 10));
-  if (parts.some(isNaN)) return 0;
-
-  let hours = 0, minutes = 0, seconds = 0;
-  if (parts.length === 3) {
-    [hours, minutes, seconds] = parts;
-  } else if (parts.length === 2) {
-    [minutes, seconds] = parts;
-  } else {
-    return 0;
-  }
-
-  return ((days * 86400) + (hours * 3600) + (minutes * 60) + seconds) * 1000;
-}
+export { parseEtime } from "@blackbelt-technology/pi-dashboard-shared/platform/process-scan.js";
+import { parseEtime } from "@blackbelt-technology/pi-dashboard-shared/platform/process-scan.js";
 
 const DEFAULT_MIN_ELAPSED_MS = 30_000;
 
@@ -112,7 +123,8 @@ export function captureChildPgids(
   trackedPgids: Set<number>,
   options?: ScanOptions,
 ): void {
-  if ((options as any)?._platform === "win32" || (!((options as any)?._platform) && process.platform === "win32")) return;
+  const platform = (options as any)?._platform ?? process.platform;
+  if (platform === "win32") return;
 
   const spawnSync: SpawnSyncFn = options?._spawnSync ?? defaultSpawnSync;
 
@@ -244,7 +256,9 @@ export function killProcessByPgid(pgid: number, options?: ScanOptions): boolean
     return killWindowsProcess(pgid, options);
   }
   try {
-    process.kill(-pgid, "SIGTERM");
+    // Route through the platform helper so the pid → -pgid mapping stays
+    // in one place. See change: route-kill-paths-through-platform.
+    killPidWithGroup(pgid, "SIGTERM", { platform });
     return true;
   } catch {
     return false;
@@ -282,9 +296,20 @@ function wmicDateToElapsedMs(creationDate: string): number {
 function getWindowsDescendants(parentPid: number, spawnSync: SpawnSyncFn): ChildProcessInfo[] {
   try {
     const result = spawnSync(
-      "wmic",
+      resolveSystemTool("wmic"),
       ["process", "where", `ParentProcessId=${parentPid}`, "get", "CommandLine,CreationDate,ParentProcessId,ProcessId", "/format:list"],
-      { encoding: "utf-8", timeout: 5000, stdio: ["pipe", "pipe", "pipe"] }
+      {
+        encoding: "utf-8",
+        timeout: 5000,
+        stdio: ["pipe", "pipe", "pipe"],
+        // Critical: without this, every 10-second process scan flashes a
+        // console window. wmic is deprecated on Win 11 but still the
+        // default here; PowerShell fallback also needs windowsHide.
+        // Also: `resolveSystemTool` above returns the FULL .exe path
+        // when the registry is available, so there's no PATHEXT /
+        // cmd.exe resolution layer that could leak a console.
+        windowsHide: true,
+      },
     );
     if (result.status !== 0 || !result.stdout) {
       // wmic removed in newer Windows 11 — fallback to tasklist
@@ -336,9 +361,17 @@ function getWindowsDescendantsTasklist(parentPid: number, spawnSync: SpawnSyncFn
     // tasklist /FI filters by parent — but tasklist doesn't support ParentProcessId filter
     // Use PowerShell Get-CimInstance as fallback
     const result = spawnSync(
-      "powershell",
+      resolveSystemTool("powershell"),
       ["-NoProfile", "-Command", `Get-CimInstance Win32_Process -Filter "ParentProcessId=${parentPid}" | Select-Object ProcessId,CommandLine,CreationDate | ConvertTo-Json`],
-      { encoding: "utf-8", timeout: 10000, stdio: ["pipe", "pipe", "pipe"] }
+      {
+        encoding: "utf-8",
+        timeout: 10000,
+        stdio: ["pipe", "pipe", "pipe"],
+        // Suppress console flash for the PowerShell fallback path.
+        // `resolveSystemTool` returns the full .exe path when registry
+        // is available.
+        windowsHide: true,
+      },
     );
     if (result.status !== 0 || !result.stdout) return [];
 
@@ -384,7 +417,8 @@ export function scanWindowsProcesses(
 export function killWindowsProcess(pid: number, options?: ScanOptions): boolean {
   const spawnSync: SpawnSyncFn = options?._spawnSync ?? defaultSpawnSync;
   try {
-    const result = spawnSync("taskkill", ["/PID", String(pid), "/T", "/F"], {
+    const result = spawnSync(resolveSystemTool("taskkill"), ["/PID", String(pid), "/T", "/F"], {
+      windowsHide: true,
       encoding: "utf-8",
       timeout: 5000,
       stdio: ["pipe", "pipe", "pipe"],

package/packages/extension/src/prompt-expander.ts

@@ -56,13 +56,20 @@ function readTemplate(filePath: string): string {
 /**
  * Expand a slash command by finding and reading the prompt template from disk.
  * Returns the expanded text, or the original text if no template found.
+ *
+ * @param pi Optional pi extension API — used to find globally installed skills
+ *           and package skills via pi.getCommands() when local scan misses them.
  */
-export function expandPromptTemplateFromDisk(text: string, cwd: string): string {
+export function expandPromptTemplateFromDisk(text: string, cwd: string, pi?: any): string {
   if (!text.startsWith("/")) return text;
 
-  const spaceIndex = text.indexOf(" ");
-  const templateName = spaceIndex === -1 ? text.slice(1) : text.slice(1, spaceIndex);
-  const argsString = spaceIndex === -1 ? "" : text.slice(spaceIndex + 1);
+  // Split template name from args on first whitespace (space OR newline).
+  // Using indexOf(" ") alone breaks multi-line payloads like "/skill:foo\nargs"
+  // because the first space can lie inside the args, producing a name such as
+  // "skill:foo\nargs-first-word" that never matches a template.
+  const m = text.slice(1).match(/^(\S+)\s*([\s\S]*)$/);
+  const templateName = m?.[1] ?? text.slice(1);
+  const argsString = m?.[2] ?? "";
 
   const templates = findPromptTemplates(cwd);
   let filePath = templates.get(templateName);
@@ -72,6 +79,20 @@ export function expandPromptTemplateFromDisk(text: string, cwd: string): string
     filePath = templates.get(templateName.replace(/:/g, "-"));
   }
 
+  // Fallback: check pi.getCommands() for globally installed skills and package skills
+  // that aren't in the local .pi/skills/ directory.
+  if (!filePath && pi?.getCommands) {
+    try {
+      const commands = pi.getCommands();
+      const skill = commands.find(
+        (c: any) => c.name === templateName && c.source === "skill" && c.path,
+      );
+      if (skill?.path && existsSync(skill.path)) {
+        filePath = skill.path;
+      }
+    } catch { /* ignore */ }
+  }
+
   if (!filePath) return text;
 
   try {

package/packages/extension/src/provider-register.ts

@@ -24,16 +24,166 @@ interface ProviderEntry {
   api?: string;
 }
 
+type InputModality = "text" | "image";
+
+export interface ModelMetadata {
+  contextWindow: number;
+  maxTokens: number;
+  reasoning: boolean;
+  cost: { input: number; output: number; cacheRead: number; cacheWrite: number };
+  input: InputModality[];
+}
+
+/**
+ * A catalog probe: given (provider, id), return the catalog entry or null.
+ * In production this is `modelRegistry.find(provider, id)` from pi's
+ * ModelRegistry (which knows both built-in pi-ai models AND user-configured
+ * custom models). Exposed as a parameter so unit tests can supply a fake
+ * catalog without needing pi-ai installed.
+ */
+export type CatalogProbe = (provider: string, modelId: string) => {
+  contextWindow: number;
+  maxTokens: number;
+  reasoning: boolean;
+  cost: { input: number; output: number; cacheRead: number; cacheWrite: number };
+  input: readonly ("text" | "image")[];
+} | null | undefined;
+
+// -- Model metadata enrichment --------------------------------------------
+//
+// Custom-provider `/v1/models` endpoints return only { id, object, ... } —
+// they do not advertise context_window, max_tokens, cost, or reasoning
+// capability. Rather than hardcode 200k / 16k / $0 / no-reasoning for every
+// discovered model (the prior behavior, which was wrong for Opus 4.6+/Sonnet
+// 4.6+/GPT-5/Gemini 2.5), we consult pi's `modelRegistry.find(provider, id)`
+// — which surfaces pi-ai's bundled catalog plus any custom models — for
+// accurate metadata and fall back to api-appropriate defaults when the
+// catalog has no match.
+//
+// See change: enrich-custom-provider-model-metadata.
+
+// API type → ordered list of candidate providers in pi's catalog.
+// Provider keys match pi-ai's MODELS export as surfaced by modelRegistry.
+// Order matters: first match wins.
+const CANDIDATE_PROVIDERS: Record<string, readonly string[]> = {
+  "anthropic-messages": ["anthropic", "opencode"],
+  "google-generative-ai": ["google", "google-vertex"],
+  "openai-completions": ["openai", "openrouter", "groq", "xai", "mistral"],
+};
+
+// Api-typed fallback defaults when the catalog has no match. Modern floors:
+//   - anthropic-messages: 200k ctx (Claude 3/4 floor), 64k maxTok
+//   - google-generative-ai: 1M ctx (Gemini 1.5+/2.x floor), 65k maxTok
+//   - openai-completions (default): 128k ctx (GPT-4o floor), 16k maxTok
+const FALLBACK_DEFAULTS: Record<string, Omit<ModelMetadata, "input">> = {
+  "anthropic-messages": {
+    contextWindow: 200_000,
+    maxTokens: 64_000,
+    reasoning: false,
+    cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+  },
+  "google-generative-ai": {
+    contextWindow: 1_000_000,
+    maxTokens: 65_536,
+    reasoning: false,
+    cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+  },
+  "openai-completions": {
+    contextWindow: 128_000,
+    maxTokens: 16_384,
+    reasoning: false,
+    cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+  },
+};
+
+const DEFAULT_INPUT: InputModality[] = ["text", "image"];
+
+/**
+ * Resolve a discovered custom-provider model id to full metadata by consulting
+ * pi's model catalog via the supplied `probe` function. Falls back to
+ * api-appropriate defaults when no catalog entry matches OR when no probe is
+ * available (e.g., modelRegistry not yet captured from spawn-context).
+ *
+ * Strips common proxy-prefix path segments (`cc/`, `anthropic/`,
+ * `openrouter/anthropic/…`) before lookup so prefixed ids resolve to the same
+ * catalog entry as the bare id.
+ *
+ * Exported (with the `probe` parameter) for unit testing. Production callers
+ * use `registerEntry()` which injects `modelRegistry.find`.
+ */
+export function enrichModelMetadata(
+  discoveredId: string,
+  api?: string,
+  probe?: CatalogProbe | null,
+): ModelMetadata {
+  const resolvedApi = api && api in CANDIDATE_PROVIDERS ? api : "openai-completions";
+  const candidates = CANDIDATE_PROVIDERS[resolvedApi] ?? CANDIDATE_PROVIDERS["openai-completions"];
+
+  // Build dedup'd list of ids to try: full, then everything after the last `/`.
+  const lookupIds: string[] = [discoveredId];
+  const lastSlash = discoveredId.lastIndexOf("/");
+  if (lastSlash >= 0 && lastSlash < discoveredId.length - 1) {
+    const bare = discoveredId.slice(lastSlash + 1);
+    if (bare && bare !== discoveredId) lookupIds.push(bare);
+  }
+
+  if (probe) {
+    for (const id of lookupIds) {
+      for (const provider of candidates) {
+        let match: ReturnType<CatalogProbe> | undefined;
+        try {
+          match = probe(provider, id);
+        } catch {
+          match = undefined;
+        }
+        if (match) {
+          return {
+            contextWindow: match.contextWindow,
+            maxTokens: match.maxTokens,
+            reasoning: match.reasoning,
+            cost: match.cost,
+            input: [...match.input] as InputModality[],
+          };
+        }
+      }
+    }
+  }
+
+  // No probe, or no catalog match — use api-appropriate fallback with
+  // image-capable default (see change: enable-image-input-custom-providers).
+  const fallback = FALLBACK_DEFAULTS[resolvedApi] ?? FALLBACK_DEFAULTS["openai-completions"];
+  return {
+    ...fallback,
+    input: [...DEFAULT_INPUT],
+  };
+}
+
 // -- Config path ----------------------------------------------------------
 
-const CONFIG_PATH = join(homedir(), ".pi", "agent", "providers.json");
+// Resolved lazily so HOME can be changed in tests.
+function configPath(): string {
+  return join(homedir(), ".pi", "agent", "providers.json");
+}
+const CONFIG_PATH = configPath();
+
+// Snapshot of last-registered provider entries so reloadProviders can diff.
+const lastRegistered = new Map<string, ProviderEntry>();
+
+function entriesEqual(a: ProviderEntry, b: ProviderEntry): boolean {
+  return (
+    a.baseUrl === b.baseUrl &&
+    a.apiKey === b.apiKey &&
+    (a.api ?? "openai-completions") === (b.api ?? "openai-completions")
+  );
+}
 
 // -- Config I/O (read-only — providers section) ----------------------------
 
 function loadProviders(): Record<string, ProviderEntry> {
-  if (existsSync(CONFIG_PATH)) {
+  const path = configPath();
+  if (existsSync(path)) {
     try {
-      const raw = JSON.parse(readFileSync(CONFIG_PATH, "utf-8"));
+      const raw = JSON.parse(readFileSync(path, "utf-8"));
       const providers: Record<string, ProviderEntry> = { ...raw.providers };
       for (const [, entry] of Object.entries(providers) as [string, any][]) {
         if (entry.apiKeyEnv && !entry.apiKey) {
@@ -43,8 +193,10 @@ function loadProviders(): Record<string, ProviderEntry> {
         delete (entry as any).modelIds;
       }
       return providers;
-    } catch {
-      // Fall through to empty
+    } catch (err: any) {
+      console.error(
+        `[dashboard] providers.json reload failed: ${err?.message ?? String(err)}`,
+      );
     }
   }
   return {};
@@ -124,6 +276,11 @@ let currentSessionModelId = "";
 
 let piRef: ExtensionAPI | null = null;
 
+// Captured from any pi event handler's ctx.modelRegistry (first available wins).
+// Used by getModelRegistry() to probe pi's catalog for model metadata enrichment.
+// See change: enrich-custom-provider-model-metadata.
+let modelRegistryRef: any = null;
+
 // Callback for notifying the bridge when providers change
 let onProvidersChanged: (() => void) | null = null;
 
@@ -148,13 +305,15 @@ export function onProviderChanged(callback: () => void): void {
   onProvidersChanged = callback;
 }
 
-// -- Helper: get modelRegistry via event ----------------------------------
-
+// -- Helper: get modelRegistry --------------------------------------------
+//
+// pi's ModelRegistry is passed as `ctx.modelRegistry` to every extension
+// event handler (see ExtensionContext in pi's types). We capture the first
+// reference we see in `session_start` and reuse it thereafter. This avoids
+// depending on pi-flows' `flow:get-spawn-context` event which is not
+// guaranteed to be present in every install.
 function getModelRegistry(): any {
-  if (!piRef) return null;
-  const spawnCtx: any = {};
-  piRef.events.emit("flow:get-spawn-context", spawnCtx);
-  return spawnCtx.modelRegistry ?? null;
+  return modelRegistryRef;
 }
 
 // -- Provider registration (with auto-discovery) --------------------------
@@ -162,14 +321,23 @@ function getModelRegistry(): any {
 async function registerEntry(pi: ExtensionAPI, name: string, entry: ProviderEntry): Promise<number> {
   const discovered = await discoverModels(entry.baseUrl, entry.apiKey);
 
+  // Metadata (contextWindow, maxTokens, reasoning, cost, input) is resolved
+  // via pi's `modelRegistry.find(provider, id)` when the registry is
+  // reachable, with api-appropriate fallbacks otherwise — the previous
+  // hardcoded 200k / 16k / $0 / no-reasoning was silently wrong for
+  // Opus 4.6+/Sonnet 4.6+/GPT-5/Gemini-2.x proxied via OpenAI-compatible
+  // endpoints. See enrichModelMetadata above, and change:
+  // enrich-custom-provider-model-metadata.
+  const registry = getModelRegistry();
+  const probe: CatalogProbe | null =
+    registry && typeof registry.find === "function"
+      ? (provider, modelId) => registry.find(provider, modelId) ?? null
+      : null;
+
   const models = discovered.map((m) => ({
     id: m.id,
     name: m.id,
-    reasoning: false,
-    input: ["text"] as ("text" | "image")[],
-    contextWindow: 200000,
-    maxTokens: 16384,
-    cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+    ...enrichModelMetadata(m.id, entry.api, probe),
   }));
 
   pi.registerProvider(name, {
@@ -179,12 +347,83 @@ async function registerEntry(pi: ExtensionAPI, name: string, entry: ProviderEntr
     models,
   });
 
+  // Record snapshot so reloadProviders can detect subsequent changes.
+  lastRegistered.set(name, {
+    baseUrl: entry.baseUrl,
+    apiKey: entry.apiKey,
+    api: entry.api ?? "openai-completions",
+  });
+
   // Notify bridge directly (same package — no cross-package event needed)
   onProvidersChanged?.();
 
   return discovered.length;
 }
 
+/**
+ * Diff the current providers.json against the last-registered snapshot and
+ * apply add / remove / change operations via `pi.registerProvider` and
+ * `pi.unregisterProvider`. Called by the bridge's `credentials_updated`
+ * handler so adding/editing/removing providers in the dashboard UI takes
+ * effect without a session restart.
+ *
+ * Malformed providers.json or IO errors produce an empty diff and do not
+ * throw, so the caller can still run `modelRegistry.refresh()` for other
+ * credential updates.
+ */
+export async function reloadProviders(
+  pi: ExtensionAPI,
+): Promise<{ added: string[]; removed: string[]; changed: string[] }> {
+  piRef = pi;
+  const added: string[] = [];
+  const removed: string[] = [];
+  const changed: string[] = [];
+
+  let current: Record<string, ProviderEntry>;
+  try {
+    current = loadProviders();
+  } catch {
+    return { added, removed, changed };
+  }
+
+  // Detect removals and changes against previous snapshot.
+  for (const [name, prev] of lastRegistered) {
+    const next = current[name];
+    if (!next) {
+      try {
+        pi.unregisterProvider(name);
+      } catch (err: any) {
+        console.error(`[dashboard] unregisterProvider("${name}") failed: ${err?.message ?? String(err)}`);
+      }
+      lastRegistered.delete(name);
+      removed.push(name);
+    } else if (!entriesEqual(prev, next)) {
+      try {
+        pi.unregisterProvider(name);
+      } catch (err: any) {
+        console.error(`[dashboard] unregisterProvider("${name}") failed: ${err?.message ?? String(err)}`);
+      }
+      lastRegistered.delete(name);
+      changed.push(name);
+    }
+  }
+
+  // Register new entries and changed entries (order-dependent: unregister ran first above).
+  for (const [name, entry] of Object.entries(current)) {
+    if (lastRegistered.has(name)) continue;
+    try {
+      await registerEntry(pi, name, entry);
+      if (!added.includes(name) && !changed.includes(name)) {
+        added.push(name);
+      }
+    } catch (err: any) {
+      console.error(`[dashboard] registerProvider("${name}") failed: ${err?.message ?? String(err)}`);
+    }
+  }
+
+  return { added, removed, changed };
+}
+
 // -- Extension entry point ------------------------------------------------
 
 export function activate(pi: ExtensionAPI) {
@@ -248,6 +487,11 @@ export function activate(pi: ExtensionAPI) {
   // ── Session lifecycle ──────────────────────────────────────────────
 
   pi.on("model_select", async (_event, ctx) => {
+    // Also capture modelRegistry here as a belt-and-suspenders in case
+    // session_start ran before activate() finished in some edge case.
+    if (!modelRegistryRef && ctx.modelRegistry) {
+      modelRegistryRef = ctx.modelRegistry;
+    }
     if (ctx.model) {
       currentSessionProvider = ctx.model.provider ?? "";
       currentSessionModelId = ctx.model.id ?? "";
@@ -255,6 +499,50 @@ export function activate(pi: ExtensionAPI) {
   });
 
   pi.on("session_start", async (_event, ctx) => {
+    // Capture the modelRegistry reference the first time we see it, then
+    // re-register already-registered providers so their model metadata gets
+    // enriched from pi's catalog (they were registered at activate() before
+    // any ctx was available, so they currently carry fallback defaults).
+    // See change: enrich-custom-provider-model-metadata.
+    if (!modelRegistryRef && ctx.modelRegistry) {
+      modelRegistryRef = ctx.modelRegistry;
+      if (lastRegistered.size > 0) {
+        // Force re-registration: clear snapshot so reloadProviders re-adds all
+        // entries (which will now probe the captured registry).
+        const names = Array.from(lastRegistered.keys());
+        lastRegistered.clear();
+        for (const name of names) {
+          const entry = providers[name];
+          if (entry) {
+            try {
+              await registerEntry(pi, name, entry);
+            } catch (err: any) {
+              console.error(`[dashboard] re-registerProvider("${name}") failed: ${err?.message ?? String(err)}`);
+            }
+          }
+        }
+
+        // If the session's currently-selected model belongs to one of the
+        // providers we just re-registered, re-apply it via pi.setModel() so
+        // the snapshot on agent.state.model picks up the enriched metadata
+        // (reasoning / contextWindow / cost). Without this, pi's session
+        // still holds the pre-enrichment descriptor with reasoning: false,
+        // causing setThinkingLevel to clamp to "off" even though the registry
+        // now has reasoning: true. See change: enrich-custom-provider-model-metadata.
+        const current = ctx.model as any;
+        if (current?.provider && current?.id && names.includes(current.provider)) {
+          try {
+            const refreshed = ctx.modelRegistry.find(current.provider, current.id);
+            if (refreshed && (pi as any).setModel) {
+              await (pi as any).setModel(refreshed);
+            }
+          } catch (err: any) {
+            console.error(`[dashboard] re-setModel after enrichment failed: ${err?.message ?? String(err)}`);
+          }
+        }
+      }
+    }
+
     if (ctx.model) {
       currentSessionProvider = ctx.model.provider ?? "";
       currentSessionModelId = ctx.model.id ?? "";