pi-better-subagents 0.1.12 → 0.1.14

package/batch.mjs

@@ -9,6 +9,7 @@
 
 import { BUILTIN_TOOLS } from "./extensions.mjs";
 import { SAFE_CLEAN_TOOLS } from "./config.ts";
+import { assertThinkingLevel } from "./thinking.ts";
 
 const VALID_CAPACITY_MODES = new Set(["reject", "launch-available"]);
 
@@ -21,6 +22,7 @@ export function mergeJobOptions(shared, job) {
         prompt: job.prompt,
         name: job.name ?? shared?.name,
         model: job.model ?? shared?.model,
+        thinking: job.thinking ?? shared?.thinking,
         tools: job.tools ?? shared?.tools,
         exclude_tools: job.exclude_tools ?? shared?.exclude_tools,
         clean: job.clean ?? shared?.clean,
@@ -74,6 +76,7 @@ export function validateBatchPlan({ shared, jobs, onCapacity, config }) {
         seenPrompts.add(job.prompt);
 
         const merged = mergeJobOptions(shared, job);
+        assertThinkingLevel(merged.thinking, `job ${i + 1} thinking`);
         if (merged.clean === true) {
             // Validate the effective tool allowlist, which resolves in this order:
             // per-job tools → shared tools → config.defaultTools → clean-safe built-ins.

package/index.ts

@@ -106,6 +106,7 @@ import {
     buildNavigatorDetail,
 } from "./navigator.ts";
 import { enforceRegistrySizeCapOnce, runDailyCleanupOnce } from "./cleanup.ts";
+import { assertThinkingLevel, parseModelThinking, type ThinkingLevel } from "./thinking.ts";
 
 /** The tools this extension registers — excluded from children by default so a
  *  subagent cannot recursively spawn more subagents unless explicitly allowed. */
@@ -613,11 +614,7 @@ function navigatorRows(visible?: RunMeta[], at?: number) {
             const snap = spendFor(m.id, now);
             return snap.tool ?? "";
         },
-        // Effort is shown when available on metadata; Pi does not always expose it.
-        effortFor: (m: RunMeta) => {
-            const any = m as RunMeta & { effort?: string; modelEffort?: string };
-            return any.effort ?? any.modelEffort;
-        },
+        effortFor: (m: RunMeta) => m.effort,
         healthFor: (m: RunMeta) => observeNavigatorHealth(m, now),
     });
 }
@@ -668,10 +665,7 @@ function navigatorDetail(id: string, now: number = Date.now()) {
         fmtElapsed,
         fmtSpend,
         now,
-        effortFor: (m: RunMeta) => {
-            const any = m as RunMeta & { effort?: string; modelEffort?: string };
-            return any.effort ?? any.modelEffort;
-        },
+        effortFor: (m: RunMeta) => m.effort,
         healthFor: (m: RunMeta) => observeNavigatorHealth(m, now),
     });
 }
@@ -917,7 +911,7 @@ export default function (pi: ExtensionAPI) {
     ensureSubagentProvider();
 
     type SpawnParams = {
-        prompt: string; name?: string; model?: string; tools?: string;
+        prompt: string; name?: string; model?: string; thinking?: ThinkingLevel; tools?: string;
         exclude_tools?: string; clean?: boolean; sandbox?: boolean;
         sandbox_dir?: string; callback?: boolean; cwd?: string;
         git_clone_workspace?: boolean; approve?: boolean; allow_nested?: boolean;
@@ -941,7 +935,10 @@ export default function (pi: ExtensionAPI) {
         warn: string;
         sandboxDir?: string;
     }> {
+        assertThinkingLevel(p.thinking);
         const cfg = loadConfig();
+        const requestedModel = p.model ?? cfg.defaultModel ?? (ctx.model ? `${ctx.model.provider}/${ctx.model.id}` : undefined);
+        const { model, thinking } = parseModelThinking(requestedModel, p.thinking);
         // Best-effort daily hygiene for durable tmp state. The marker makes
         // this effectively free after the first subagent launch each day.
         runDailyCleanupOnce({ config: cfg });
@@ -978,7 +975,6 @@ export default function (pi: ExtensionAPI) {
         });
         const cwd = workspace.cwd;
         const requestedSandboxDir = workspace.requestedSandboxDir;
-        const model = p.model ?? cfg.defaultModel ?? (ctx.model ? `${ctx.model.provider}/${ctx.model.id}` : undefined);
 
         if (requestedSandboxDir) mkdirSync(requestedSandboxDir, { recursive: true });
         writeFileSync(promptPathFor(id), p.prompt);
@@ -1017,6 +1013,7 @@ export default function (pi: ExtensionAPI) {
             "--session-id", id,
             ...extArgs,
             ...(model ? ["--model", model] : []),
+            ...(thinking ? ["--thinking", thinking] : []),
             ...(allow ? ["--tools", allow] : []),
             ...(excludes.size ? ["--exclude-tools", [...excludes].join(",")] : []),
             ...(p.approve ? ["--approve"] : []),
@@ -1043,7 +1040,8 @@ export default function (pi: ExtensionAPI) {
 
         const meta: RunMeta = {
             id, name: p.name, status: "running",
-            pid: spawned.pid, spawnPid: process.pid, spawnPidStartTime: parentStartToken(), model, cwd,
+            pid: spawned.pid, spawnPid: process.pid, spawnPidStartTime: parentStartToken(), model,
+            effort: thinking, cwd,
             ...identity,
             promptPreview: p.prompt.slice(0, 200),
             startedAt: Date.now(), logPath: logPathFor(id), sessionId: id,
@@ -1088,14 +1086,15 @@ export default function (pi: ExtensionAPI) {
             "Use subagent_spawn for independent work the user should not have to wait on. It returns at once with a run id; that return IS the deliverable — report the id to the user and continue.",
             "After subagent_spawn, do NOT call subagent_output or subagent_result in a loop to wait for the result, and do NOT sleep. The run completes on its own and reports back on the next turn.",
             "Only call subagent_result / subagent_output when the user explicitly asks how a run is going or for its result.",
-            "The tools param is both the tool allowlist AND what determines which extensions load in the child (e.g. tools='read,bash,web_fetch' loads only the web-tools package). Ask for the tools the task needs and nothing more; clean:true gives a built-ins-only child. Pick a model with the model param (e.g. 'xai/grok-4.5').",
+            "The tools param is both the tool allowlist AND what determines which extensions load in the child (e.g. tools='read,bash,web_fetch' loads only the web-tools package). Ask for the tools the task needs and nothing more; clean:true gives a built-ins-only child. Pick a model with the model param (e.g. 'xai/grok-4.5@high'); providerless model patterns are resolved by Pi, while provider/model is deterministic and loads mapped provider extensions.",
             "By default the subagent is sandboxed (writes confined to its working dir, reads and network open) and triggers completion here on finish. Set callback:false to finish quietly — then read the result on demand via subagent_result.",
             "Use git_clone_workspace:true when the subagent will mutate Git in a sandbox. The parent prepares a disposable, self-contained clone with a real .git/ directory inside the sandbox root, so linked-worktree metadata outside the sandbox cannot stall the child.",
         ],
         parameters: Type.Object({
             prompt: Type.String({ description: "The task for the subagent. This is the only context it gets — be self-contained." }),
             name: Type.Optional(Type.String({ description: "Short label for the run (e.g. 'reviewer')." })),
-            model: Type.Optional(Type.String({ description: "Model as provider/id (default: inherit foreground model)." })),
+            model: Type.Optional(Type.String({ description: "Pi model pattern, preferably provider/id, optionally suffixed with @effort (for example openai/gpt-5.5@high). Providerless patterns are resolved by Pi. Default: inherit foreground model." })),
+            thinking: Type.Optional(Type.String({ description: "Reasoning effort for the child: off, minimal, low, medium, high, xhigh, or max (default: Pi/model default)." })),
             tools: Type.Optional(Type.String({ description: "Tool allowlist: comma-separated names the child may use (e.g. 'read,bash,web_fetch'). This ALSO selects which extensions load — only packages backing a requested tool are loaded. Defaults to the configured safe set." })),
             exclude_tools: Type.Optional(Type.String({ description: "Comma-separated tool denylist, applied on top of the allowlist." })),
             clean: Type.Optional(Type.Boolean({ description: "Run a hermetic child with NO extensions at all (only built-ins: read, bash, edit, write). Default false — the extensions backing the requested tools load, so web_fetch and model auth (e.g. xai) work." })),
@@ -1160,7 +1159,8 @@ export default function (pi: ExtensionAPI) {
         parameters: Type.Object({
             batchName: Type.Optional(Type.String({ description: "Optional display label for the batch." })),
             shared: Type.Optional(Type.Object({
-                model: Type.Optional(Type.String({ description: "Model as provider/id (default: inherit foreground model)." })),
+                model: Type.Optional(Type.String({ description: "Pi model pattern, preferably provider/id, optionally suffixed with @effort (default: inherit foreground model)." })),
+                thinking: Type.Optional(Type.String({ description: "Reasoning effort applied to every job: off, minimal, low, medium, high, xhigh, or max." })),
                 tools: Type.Optional(Type.String({ description: "Tool allowlist applied to every job." })),
                 exclude_tools: Type.Optional(Type.String({ description: "Comma-separated tool denylist applied to every job." })),
                 sandbox: Type.Optional(Type.Boolean({ description: "Default TRUE: kernel-confine writes to the working dir." })),
@@ -1176,6 +1176,7 @@ export default function (pi: ExtensionAPI) {
                 prompt: Type.String({ description: "The task for this job." }),
                 name: Type.Optional(Type.String({ description: "Short label for this job." })),
                 model: Type.Optional(Type.String()),
+                thinking: Type.Optional(Type.String()),
                 tools: Type.Optional(Type.String()),
                 exclude_tools: Type.Optional(Type.String()),
                 sandbox: Type.Optional(Type.Boolean()),

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-better-subagents",
-  "version": "0.1.12",
+  "version": "0.1.14",
   "description": "Pi extension for detached, sandboxed subagent runs that keep the foreground session free.",
   "license": "MIT",
   "type": "module",

package/registry.ts

@@ -85,6 +85,8 @@ export interface RunMeta {
      */
     adoptedFromLostParentAt?: number;
     model?: string;
+    /** Reasoning effort passed to the child via Pi's --thinking option. */
+    effort?: string;
     cwd: string;
     /** First ~200 chars of the task prompt, for listings. */
     promptPreview: string;

package/thinking.ts

@@ -0,0 +1,28 @@
+export const THINKING_LEVELS = ["off", "minimal", "low", "medium", "high", "xhigh", "max"] as const;
+
+export type ThinkingLevel = typeof THINKING_LEVELS[number];
+
+const VALID_THINKING_LEVELS = new Set<string>(THINKING_LEVELS);
+
+export function assertThinkingLevel(value: unknown, label: string = "thinking"): asserts value is ThinkingLevel | undefined {
+    if (value === undefined || VALID_THINKING_LEVELS.has(String(value))) return;
+    throw new Error(`${label} must be one of: ${THINKING_LEVELS.join(", ")}; got ${JSON.stringify(value)}.`);
+}
+
+/** Resolve the convenient `model@effort` shorthand used by subagent callers. */
+export function parseModelThinking(
+    model: string | undefined,
+    thinking?: ThinkingLevel,
+): { model: string | undefined; thinking: ThinkingLevel | undefined } {
+    if (!model) return { model, thinking };
+
+    const separator = model.lastIndexOf("@");
+    if (separator <= 0 || separator === model.length - 1) return { model, thinking };
+
+    const suffix = model.slice(separator + 1);
+    assertThinkingLevel(suffix, "model thinking suffix");
+    return {
+        model: model.slice(0, separator),
+        thinking: thinking ?? suffix,
+    };
+}