goalbuddy 0.3.5 → 0.3.7

package/goalbuddy/scripts/check-goal-state.mjs

@@ -53,6 +53,38 @@ function nestedScalar(section, key) {
   return null;
 }
 
+function pathScalar(path, key) {
+  const lines = text.split(/\r?\n/);
+  let depth = 0;
+  for (const line of lines) {
+    if (!line.trim()) continue;
+    const indent = line.match(/^ */)[0].length;
+    if (indent < depth * 2) depth = Math.floor(indent / 2);
+
+    if (depth < path.length && indent === depth * 2 && new RegExp(`^\\s{${indent}}${path[depth]}:\\s*$`).test(line)) {
+      depth += 1;
+      continue;
+    }
+
+    if (depth === path.length && indent === depth * 2) {
+      const match = line.match(new RegExp(`^\\s{${indent}}${key}:\\s*(.*?)\\s*$`));
+      if (match) return clean(match[1]);
+    }
+  }
+  return null;
+}
+
+function isWeakProof(value) {
+  if (value === null || value === undefined) return true;
+  const normalized = String(value).trim().toLowerCase();
+  return normalized === ""
+    || normalized === "unknown"
+    || normalized === "tbd"
+    || normalized === "todo"
+    || normalized === "none"
+    || /^<.*>$/.test(normalized);
+}
+
 function sectionText(section) {
   const lines = text.split(/\r?\n/);
   const start = lines.findIndex((line) => new RegExp(`^${section}:\\s*$`).test(line));
@@ -225,6 +257,11 @@ const allowedAgentStatuses = new Set(["installed", "bundled_not_installed", "mis
 const continuousUntilFullOutcome = nestedScalar("rules", "continuous_until_full_outcome") === true;
 const missingInputOrCredentialsDoNotStopGoal =
   nestedScalar("rules", "missing_input_or_credentials_do_not_stop_goal") === true;
+const goalPressureRequiresOracle = nestedScalar("rules", "goal_pressure_requires_oracle") !== false;
+const noCompletionOnWeakProof = nestedScalar("rules", "no_completion_on_weak_proof") !== false;
+const completionProof = pathScalar(["goal", "intake"], "completion_proof");
+const oracleSignal = pathScalar(["goal", "oracle"], "signal");
+const oracleFinalProof = pathScalar(["goal", "oracle"], "final_proof");
 const legacySignals = [
   /^gate:\s*$/m,
   /^artifact_policy:\s*$/m,
@@ -244,6 +281,19 @@ if (!["active", "blocked", "done"].includes(goalStatus)) {
   errors.push(`goal.status must be active, blocked, or done; got ${goalStatus || "<missing>"}`);
 }
 
+if (goalPressureRequiresOracle) {
+  if (isWeakProof(oracleSignal)) {
+    warnings.push("goal.oracle.signal is missing or placeholder-like; weak oracles make /goal finish too early.");
+  }
+  if (isWeakProof(oracleFinalProof)) {
+    warnings.push("goal.oracle.final_proof is missing or placeholder-like; final completion needs receipt-backed proof.");
+  }
+}
+
+if (isWeakProof(completionProof)) {
+  warnings.push("goal.intake.completion_proof is missing or placeholder-like; record the observable signal that proves the full original outcome.");
+}
+
 function agentStatusWarning(agent, status) {
   const agentLabel = agent[0].toUpperCase() + agent.slice(1);
   if (status === "bundled_not_installed") {
@@ -369,6 +419,9 @@ for (const task of tasks) {
         errors.push(`Worker receipt for ${task.id} has non-passing command status: ${status}`);
       }
     }
+    if (task.receipt.scalar("needs_judge") === true) {
+      warnings.push(`Worker receipt for ${task.id} requests legacy needs_judge; GoalBuddy now lets the PM continue by default and reviews only at phase, risk, ambiguity, rejected-verification, or final-completion boundaries`);
+    }
   }
   if (task.type === "scout" && task.status === "done" && hasReceipt) {
     if (!task.receipt.has("summary")) errors.push(`Scout receipt for ${task.id} missing summary`);
@@ -381,6 +434,8 @@ for (const task of tasks) {
   }
 }
 
+warnings.push(...microSliceWarnings(tasks, activeTask, goalStatus));
+
 function validateSubgoal(task) {
   if (isChildCheck) {
     errors.push(`child task ${task.id} must not contain a nested subgoal`);
@@ -430,6 +485,77 @@ function validateSubgoal(task) {
   }
 }
 
+function microSliceWarnings(tasks, activeTaskId, goalStatus) {
+  const found = [];
+  const guidance = "Board may be micro-slicing. Prefer the largest safe useful slice.";
+  const doneTasks = tasks.filter((task) => task.status === "done");
+  const workerTasks = tasks.filter((task) => task.type === "worker");
+  const recentTinyWorkers = workerTasks.slice(-5).filter((task) => isTinyTask(task));
+  const firstMilestoneComplete = nestedScalar("goal", "first_milestone_complete") === true;
+
+  if (recentTinyWorkers.length >= 3) {
+    found.push(`${guidance} Three recent Worker tasks look tiny.`);
+  }
+
+  for (const task of tasks) {
+    if (task.type === "judge" && /pick small reviewable work|select one narrow next task/i.test(task.raw)) {
+      found.push(`${guidance} Judge instructions still ask for small or narrow work.`);
+      break;
+    }
+  }
+
+  if (goalStatus !== "active" || !activeTaskId) return [...new Set(found)];
+  const activeIndex = tasks.findIndex((task) => task.id === activeTaskId);
+  if (activeIndex === -1) return [...new Set(found)];
+  const active = tasks[activeIndex];
+  if (active.type === "worker") {
+    if (doneTasks.length >= 10 && active.allowedFiles.length > 0 && active.allowedFiles.length <= 2) {
+      found.push(`${guidance} Active Worker ${active.id} has only ${active.allowedFiles.length} allowed_files after ${doneTasks.length} completed tasks.`);
+    }
+    if (firstMilestoneComplete && isTinyTask(active)) {
+      found.push(`${guidance} The first milestone is complete, so the active Worker should move toward the next real milestone.`);
+    }
+    if (isMicroWorkerTask(active)) {
+      found.push(`${guidance} Active Worker ${active.id} looks like another helper-sized slice.`);
+    }
+  }
+  if (active.type !== "judge") return [...new Set(found)];
+
+  let pairs = 0;
+  for (let index = activeIndex; index > 0; index -= 2) {
+    const judge = tasks[index];
+    const worker = tasks[index - 1];
+    if (!isMicroJudgeForWorker(judge, worker)) break;
+    pairs += 1;
+  }
+  if (pairs >= 2) {
+    found.push(`${guidance} Micro Worker/Judge loop detected ending at ${active.id}.`);
+  }
+  return [...new Set(found)];
+}
+
+function isMicroJudgeForWorker(judge, worker) {
+  if (!judge || !worker) return false;
+  if (judge.type !== "judge" || worker.type !== "worker") return false;
+  if (!["active", "queued", "done"].includes(judge.status) || worker.status !== "done") return false;
+  const objective = String(judge.objective || "").toLowerCase();
+  return objective.includes(worker.id.toLowerCase()) && /audit|review|approve/.test(objective) && isMicroWorkerTask(worker);
+}
+
+function isMicroWorkerTask(task) {
+  if (!task || task.type !== "worker") return false;
+  const objective = String(task.objective || "").toLowerCase();
+  if (/collapsed|batch|package|tranche/.test(objective)) return false;
+  return /one narrow|single helper|one helper|per[- ]helper|per[- ]table|projection helper/.test(objective);
+}
+
+function isTinyTask(task) {
+  if (!task) return false;
+  const text = [task.objective, task.raw, task.receipt?.raw].join(" ").toLowerCase();
+  if (/collapsed|batch|package|tranche|vertical slice|milestone/.test(text)) return false;
+  return /\b(tiny|narrow|single helper|one helper|projection helper|projection function|contract file|read-only proof|doc note|validator|validation wrapper|pure helper|caller-input)\b/.test(text);
+}
+
 function matchesAllowedFile(file, allowedFiles) {
   return allowedFiles.some((pattern) => globMatch(pattern, file));
 }
@@ -456,6 +582,9 @@ function escapeRegExp(value) {
 }
 
 if (goalStatus === "done") {
+  if (noCompletionOnWeakProof && (isWeakProof(completionProof) || isWeakProof(oracleSignal) || isWeakProof(oracleFinalProof))) {
+    errors.push("done goals require concrete completion proof, goal.oracle.signal, and goal.oracle.final_proof; weak proof cannot close a goal");
+  }
   const finalAudit = tasks.some((task) => {
     if (!["judge", "pm"].includes(task.type) || task.status !== "done") return false;
     if (!task.receipt.present || task.receipt.value === null) return false;

package/goalbuddy/scripts/render-task-prompt.mjs

@@ -2,12 +2,12 @@
 import { existsSync, readFileSync } from "node:fs";
 import { basename, dirname, resolve } from "node:path";
 import { fileURLToPath } from "node:url";
-import { parseGoalStateText } from "../extend/local-goal-board/scripts/lib/goal-board.mjs";
+import { parseGoalStateText } from "../surfaces/local-goal-board/scripts/lib/goal-board.mjs";
 
 const ROLE_DEFAULTS = {
   scout: { agent: "goal_scout", reasoning: "low", sandbox: "read-only" },
   judge: { agent: "goal_judge", reasoning: "high", sandbox: "read-only" },
-  worker: { agent: "goal_worker", reasoning: "low", sandbox: "workspace-write" },
+  worker: { agent: "goal_worker", reasoning: "medium", sandbox: "workspace-write" },
   pm: { agent: "PM", reasoning: "medium", sandbox: "workspace-write" },
 };
 
@@ -39,11 +39,14 @@ export function renderTaskPrompt(options) {
     payload: {
       metadata: {
         recommended_agent: defaults.agent,
+        required_spawn_agent_type: defaults.agent === "PM" ? null : defaults.agent,
         recommended_reasoning: reasoning,
         sandbox: defaults.sandbox,
         fork_context_allowed: role !== "worker",
         board_path: board.path,
         child_board_paths: childBoardPaths(board),
+        goal_oracle: board.goal.oracle || null,
+        slice_policy: board.document.rules?.slice_policy || null,
         warnings,
       },
       task: {
@@ -139,19 +142,63 @@ function promptWarnings(board, task) {
   const warnings = [];
   const role = normalizeRole(task.type);
   if (task.id !== board.activeTask) warnings.push(`Task ${task.id} is not the active task on this board.`);
+  if (isWeakProof(board.goal.oracle?.signal)) {
+    warnings.push("goal.oracle.signal is missing or placeholder-like; keep the goal pressured by a concrete completion oracle.");
+  }
+  if (isWeakProof(board.goal.oracle?.final_proof)) {
+    warnings.push("goal.oracle.final_proof is missing or placeholder-like; do not mark the goal complete without receipt-backed proof.");
+  }
   if (role === "worker") {
     if (stringList(task.allowed_files).length === 0) warnings.push(`Worker task ${task.id} has no allowed_files.`);
     if (stringList(task.verify).length === 0) warnings.push(`Worker task ${task.id} has no verify commands.`);
     if (stringList(task.stop_if).length === 0) warnings.push(`Worker task ${task.id} has no stop_if conditions.`);
+    if (isFalse(board.goal.full_outcome_complete)) {
+      warnings.push(`full_outcome_complete is false and ${task.id} is an active Worker; do not stop after rendering or repairing the board. Execute the Worker unless a stop_if condition applies.`);
+    }
   }
   for (const candidate of board.tasks) {
     if (candidate?.subgoal && Number(candidate.subgoal.depth) !== 1) {
       warnings.push(`Task ${candidate.id} has subgoal.depth ${candidate.subgoal.depth || "<missing>"}; only depth 1 is supported.`);
     }
   }
+  warnings.push(...microSliceWarnings(board, task));
   return warnings;
 }
 
+function microSliceWarnings(board, task) {
+  const warnings = [];
+  const doneTasks = board.tasks.filter((candidate) => candidate?.status === "done");
+  const recentWorkers = board.tasks
+    .filter((candidate) => normalizeRole(candidate?.type) === "worker")
+    .slice(-5);
+  const recentTinyWorkers = recentWorkers.filter((candidate) => isTinyTask(candidate));
+  const activeRole = normalizeRole(task.type);
+  const activeAllowedFiles = stringList(task.allowed_files);
+  const firstMilestoneComplete = isTrue(board.goal.first_milestone_complete);
+  const microWarning = "Board may be micro-slicing. Prefer the largest safe useful slice.";
+
+  if (recentTinyWorkers.length >= 3) warnings.push(microWarning);
+  if (doneTasks.length >= 10 && activeRole === "worker" && activeAllowedFiles.length > 0 && activeAllowedFiles.length <= 2) {
+    warnings.push(`${microWarning} Active Worker ${task.id} has only ${activeAllowedFiles.length} allowed_files after ${doneTasks.length} completed tasks.`);
+  }
+  if (firstMilestoneComplete && activeRole === "worker" && isTinyTask(task)) {
+    warnings.push(`${microWarning} The first milestone is complete, so the active Worker should move toward the next real milestone.`);
+  }
+  if (activeRole === "judge" && /pick small reviewable work|select one narrow next task/i.test(String(task.objective || "") + "\n" + stringList(task.constraints).join("\n"))) {
+    warnings.push(`${microWarning} Judge instructions still ask for small or narrow work.`);
+  }
+  return [...new Set(warnings)];
+}
+
+function isTinyTask(task) {
+  const text = [
+    task?.objective,
+    stringList(task?.constraints).join(" "),
+    task?.receipt?.summary,
+  ].join(" ").toLowerCase();
+  return /\b(tiny|narrow|single helper|one helper|projection helper|projection function|contract file|read-only proof|doc note|validator|validation wrapper|pure helper|caller-input)\b/.test(text);
+}
+
 function normalizeRole(value) {
   const role = String(value || "pm").toLowerCase();
   return ROLE_DEFAULTS[role] ? role : "pm";
@@ -163,6 +210,25 @@ function normalizeReasoning(value, fallback) {
   return fallback;
 }
 
+function isFalse(value) {
+  return value === false || String(value).toLowerCase() === "false";
+}
+
+function isTrue(value) {
+  return value === true || String(value).toLowerCase() === "true";
+}
+
+function isWeakProof(value) {
+  if (value === null || value === undefined) return true;
+  const normalized = String(value).trim().toLowerCase();
+  return normalized === ""
+    || normalized === "unknown"
+    || normalized === "tbd"
+    || normalized === "todo"
+    || normalized === "none"
+    || /^<.*>$/.test(normalized);
+}
+
 function stringList(value) {
   return Array.isArray(value) ? value.filter((item) => item !== null && item !== undefined).map(String) : [];
 }
@@ -175,15 +241,14 @@ function receiptSchema(role) {
       commands: [{ cmd: "<command>", status: "pass | fail | not_run" }],
       summary: "<=120 words",
       remaining_blockers: [],
-      needs_judge: false,
     };
   }
   if (role === "judge") {
     return {
       result: "done | blocked",
-      decision: "approve_next | reject_next | approve_subgoal | reject_subgoal | not_complete | complete",
+      decision: "approved | rejected | approve_subgoal | reject_subgoal | not_complete | complete",
+      full_outcome_complete: false,
       evidence: [],
-      next_allowed_task: null,
       blocked_tasks: [],
       required_board_updates: [],
     };
@@ -204,6 +269,7 @@ function formatPrompt(payload) {
     "",
     "Metadata:",
     `- recommended_agent: ${payload.metadata.recommended_agent}`,
+    `- required_spawn_agent_type: ${payload.metadata.required_spawn_agent_type || "PM fallback"}`,
     `- recommended_reasoning: ${payload.metadata.recommended_reasoning}`,
     `- sandbox: ${payload.metadata.sandbox}`,
     `- fork_context_allowed: ${payload.metadata.fork_context_allowed}`,
@@ -213,12 +279,24 @@ function formatPrompt(payload) {
     lines.push("- child_board_paths:");
     for (const path of payload.metadata.child_board_paths) lines.push(`  - ${path}`);
   }
+  if (payload.metadata.goal_oracle) {
+    lines.push(`- goal_oracle: ${JSON.stringify(payload.metadata.goal_oracle)}`);
+  }
+  if (payload.metadata.slice_policy) {
+    lines.push(`- slice_policy: ${JSON.stringify(payload.metadata.slice_policy)}`);
+  }
   if (payload.metadata.warnings.length) {
     lines.push("- warnings:");
     for (const warning of payload.metadata.warnings) lines.push(`  - ${warning}`);
   }
 
   lines.push(
+    "",
+    "Spawn contract:",
+    `- Codex spawn_agent agent_type: ${payload.metadata.required_spawn_agent_type || "do not spawn; run as PM"}`,
+    "- Do not substitute generic scout, worker, or judge agents for GoalBuddy agents.",
+    "- If the required GoalBuddy agent is unavailable, stop spawning and continue as PM fallback or install agents.",
+    "- After one wait_agent timeout with no visible allowed-file changes, stop waiting and recover deterministically.",
     "",
     "Task:",
     `- id: ${payload.task.id}`,

package/{plugins/goalbuddy/skills/goalbuddy/extend → goalbuddy/surfaces}/local-goal-board/README.md

@@ -2,7 +2,7 @@
 
 Generate a small local GoalBuddy board for a goal directory and watch it update live while agents work.
 
-The extension keeps `state.yaml` authoritative. It writes static web app files into the goal directory and serves them from a local-only Node server. The browser subscribes to Server-Sent Events, so cards update as `state.yaml`, `notes/`, or linked depth-1 sub-goal state changes without a manual reload.
+The surface keeps `state.yaml` authoritative. It writes static web app files into the goal directory and serves them from a local-only Node server. The browser subscribes to Server-Sent Events, so cards update as `state.yaml`, `notes/`, or linked depth-1 sub-goal state changes without a manual reload.
 
 ## Use When
 
@@ -14,11 +14,10 @@ The extension keeps `state.yaml` authoritative. It writes static web app files i
 ## Generate And Serve
 
 ```bash
-node extend/local-goal-board/scripts/local-goal-board.mjs \
-  --goal docs/goals/<slug>
+npx goalbuddy board docs/goals/<slug>
 ```
 
-The generated app includes the bundled `assets/goalbuddy-mark.png`, so the board keeps the GoalBuddy mark after the extension is installed or copied elsewhere.
+The generated app includes the bundled `assets/goalbuddy-mark.png`, so the board keeps the GoalBuddy mark anywhere the package is installed.
 
 The command writes:
 
@@ -34,8 +33,7 @@ Then it starts or reuses the shared local board hub at `http://goalbuddy.localho
 ## Check Without A Long-Running Server
 
 ```bash
-node extend/local-goal-board/scripts/local-goal-board.mjs \
-  --goal docs/goals/<slug> \
+npx goalbuddy board docs/goals/<slug> \
   --once \
   --json
 ```
@@ -63,9 +61,9 @@ Clicking a card opens a detail modal with the task objective, status, assignee,
 ## Verification
 
 ```bash
-node --test extend/local-goal-board/test/*.test.mjs
-node extend/local-goal-board/scripts/local-goal-board.mjs \
-  --goal extend/local-goal-board/examples/sample-goal \
+node --test goalbuddy/surfaces/local-goal-board/test/*.test.mjs
+node goalbuddy/surfaces/local-goal-board/scripts/local-goal-board.mjs \
+  --goal goalbuddy/surfaces/local-goal-board/examples/sample-goal \
   --once \
   --json
 ```

package/{plugins/goalbuddy/skills/goalbuddy/extend → goalbuddy/surfaces}/local-goal-board/examples/sample-goal/state.yaml

@@ -1,8 +1,8 @@
 version: 2
 
 goal:
-  title: "Local Kanban Board Extension"
-  slug: "local-kanban-board-extension"
+  title: "Local Goal Board Surface"
+  slug: "local-goal-board-surface"
   kind: specific
   tranche: "Demonstrate local GoalBuddy board rendering."
   status: active
@@ -30,7 +30,7 @@ tasks:
     type: worker
     assignee: Worker
     status: blocked
-    objective: "Catalog and document the local board extension."
+    objective: "Catalog and document the local board surface."
     receipt:
       result: blocked
       summary: "T003 is blocked during the progressive board motion demo."
@@ -78,7 +78,7 @@ tasks:
     type: scout
     assignee: Scout
     status: done
-    objective: "List the extension launch paths."
+    objective: "List the board launch paths."
     receipt:
       result: done
       summary: "T009 completed during the progressive board motion demo."
@@ -118,7 +118,7 @@ tasks:
     type: worker
     assignee: Worker
     status: done
-    objective: "Prepare final extension packaging check."
+    objective: "Prepare final board packaging check."
     receipt:
       result: done
       summary: "T014 completed during the progressive board motion demo."

package/{plugins/goalbuddy/skills/goalbuddy/extend → goalbuddy/surfaces}/local-goal-board/examples/subgoal-parent/state.yaml

@@ -31,10 +31,10 @@ tasks:
     status: active
     objective: "Build the sub-goal board view."
     allowed_files:
-      - goalbuddy/extend/local-goal-board/scripts/lib/goal-board.mjs
-      - goalbuddy/extend/local-goal-board/test/local-goal-board.test.mjs
+      - goalbuddy/surfaces/local-goal-board/scripts/lib/goal-board.mjs
+      - goalbuddy/surfaces/local-goal-board/test/local-goal-board.test.mjs
     verify:
-      - node --test goalbuddy/extend/local-goal-board/test/local-goal-board.test.mjs
+      - node --test goalbuddy/surfaces/local-goal-board/test/local-goal-board.test.mjs
     stop_if:
       - "Need files outside allowed_files."
     subgoal:

package/goalbuddy/{extend → surfaces}/local-goal-board/examples/subgoal-parent/subgoals/T004-board-view/state.yaml

@@ -24,16 +24,16 @@ tasks:
       result: done
       summary: "Child board payload needs normal columns and task details."
       evidence:
-        - goalbuddy/extend/local-goal-board/scripts/lib/goal-board.mjs
+        - goalbuddy/surfaces/local-goal-board/scripts/lib/goal-board.mjs
   - id: T002
     type: worker
     assignee: Worker
     status: active
     objective: "Render the read-only embedded child board."
     allowed_files:
-      - goalbuddy/extend/local-goal-board/scripts/lib/goal-board.mjs
+      - goalbuddy/surfaces/local-goal-board/scripts/lib/goal-board.mjs
     verify:
-      - node --test goalbuddy/extend/local-goal-board/test/local-goal-board.test.mjs
+      - node --test goalbuddy/surfaces/local-goal-board/test/local-goal-board.test.mjs
     stop_if:
       - "Need files outside allowed_files."
     receipt: null