skyloom 1.13.6 → 1.13.8

package/src/core/agent/guard.ts

@@ -1,134 +1,134 @@
-/**
- * Anti-loop guard for the agent reasoning loop.
- *
- * Extracted from chatStreamImpl (Phase 3). Holds the per-turn heuristic state
- * (recent response texts, tool-call signatures, tool outcomes + once-only hint
- * flags) and, after each round, returns a decision: zero or more system
- * "hints" to nudge the model, and optionally a hard `stop` (assistant note +
- * a user-visible content line). The agent loop applies the decision — the guard
- * itself has no side effects, which makes every branch unit-testable.
- *
- * A fresh LoopGuard is created per turn (state must not leak across turns).
- */
-
-import type { ToolCall } from '../llm';
-import {
-  toolCallSignature,
-  textSimilarity,
-  looksLikeFailedToolResult,
-  parseToolArgs,
-  SIG_WINDOW,
-  SIG_LOOP_HINT,
-  SIG_LOOP_HARDSTOP,
-} from '../agent_helpers';
-
-/** A hard stop: record `note` as an assistant message, then surface `contentLine`. */
-export interface GuardStop {
-  note: string;
-  contentLine: string;
-}
-
-/** The guard's decision for one round. `hints` apply in order; `stop` ends the turn. */
-export interface GuardDecision {
-  hints: string[];
-  stop?: GuardStop;
-}
-
-/** Minimal shape of a tool execution result the guard inspects. */
-export interface GuardExecResult {
-  toolName: string;
-  success: boolean;
-  result: string;
-}
-
-export class LoopGuard {
-  private recentResponseTexts: string[] = [];
-  private recentToolSigs: string[] = [];
-  private recentToolOutcomes: boolean[] = [];
-  private searchCount = 0; // cumulative search/fetch calls this turn (not window-bounded)
-  private repetitionHintInjected = false;
-  private toolLoopHintInjected = false; // shared by tool-signature loop + search-storm
-  private stuckHintInjected = false;
-
-  /**
-   * Observe one completed round. Mutates internal state and returns the
-   * hints/stop decision. Evaluation order (and the shared hint flag) mirrors
-   * the original inline logic exactly.
-   */
-  observe(
-    roundContent: string,
-    toolCallsReceived: ToolCall[],
-    execResults: Array<GuardExecResult | null>
-  ): GuardDecision {
-    const hints: string[] = [];
-
-    // 1. Narration-loop: response too similar to a recent one.
-    const normalizedRound = (roundContent || '').trim();
-    if (normalizedRound && this.recentResponseTexts.length > 0) {
-      const highSim = this.recentResponseTexts.slice(-2).some(prev => textSimilarity(normalizedRound, prev) >= 0.7);
-      if (highSim && !this.repetitionHintInjected) {
-        hints.push('[Stop narrating] Your last response is highly similar to your previous one. Stop writing prose. Either: (1) emit ONLY the next tool call, or (2) output the final deliverable.');
-        this.repetitionHintInjected = true;
-      }
-    }
-    this.recentResponseTexts.push(normalizedRound);
-    if (this.recentResponseTexts.length > 3) this.recentResponseTexts.shift();
-
-    // 2. Tool-signature loop: same call repeated within the window.
-    for (const tc of toolCallsReceived) {
-      const tName = tc.function.name;
-      if (['task_done', 'list_skills', 'use_skill'].includes(tName)) continue;
-      if (['web_search', 'fetch_page', 'http_get'].includes(tName)) this.searchCount++;
-      const rawArgs = tc.function.arguments;
-      const tArgs = typeof rawArgs === 'string' ? parseToolArgs(rawArgs) : rawArgs;
-      const sig = toolCallSignature(tName, tArgs);
-      if (sig) this.recentToolSigs.push(sig);
-    }
-    if (this.recentToolSigs.length > SIG_WINDOW) {
-      this.recentToolSigs.splice(0, this.recentToolSigs.length - SIG_WINDOW);
-    }
-    if (this.recentToolSigs.length > 0) {
-      const counts = new Map<string, number>();
-      for (const s of this.recentToolSigs) counts.set(s, (counts.get(s) || 0) + 1);
-      let topSig = '';
-      let topCount = 0;
-      for (const [s, c] of counts) { if (c > topCount) { topSig = s; topCount = c; } }
-      if (topCount >= SIG_LOOP_HINT && !this.toolLoopHintInjected) {
-        hints.push(`[Tool loop] You have called \`${topSig}\` ${topCount}x in the last ${this.recentToolSigs.length} tool calls — you are iterating without converging. STOP repeating it.`);
-        this.toolLoopHintInjected = true;
-      }
-      if (topCount >= SIG_LOOP_HARDSTOP) {
-        return { hints, stop: { note: `I have repeated \`${topSig}\` ${topCount} times without converging. Stopping.`, contentLine: `\n\n[stuck] tool \`${topSig}\` repeated ${topCount}x — stopping.` } };
-      }
-    }
-
-    // 3. Stuck-loop: most/all recent tool calls failed.
-    for (const r of execResults) {
-      if (!r || r.toolName === 'task_done') continue;
-      const failed = !r.success || (typeof r.result === 'string' && looksLikeFailedToolResult(r.result));
-      this.recentToolOutcomes.push(!failed);
-      // Keep 8 so the "all recent calls failed" (>=8) hard-stop below is reachable.
-      if (this.recentToolOutcomes.length > 8) this.recentToolOutcomes.shift();
-    }
-    if (!this.stuckHintInjected && this.recentToolOutcomes.length >= 5 &&
-        this.recentToolOutcomes.filter(Boolean).length <= 1) {
-      hints.push('[Recovery hint] Your last several tool calls have mostly failed. Synthesize a partial answer from what worked or ask the user for guidance.');
-      this.stuckHintInjected = true;
-    }
-    if (this.recentToolOutcomes.length >= 8 && this.recentToolOutcomes.every(x => !x)) {
-      return { hints, stop: { note: 'Every recent tool call failed. Please give me more context.', contentLine: '\n\n[stuck] every recent tool call failed — stopping.\n' } };
-    }
-
-    // 4. Search-storm: cumulative search/fetch calls this turn (not bounded by
-    // SIG_WINDOW, so the >=12 hard-stop is actually reachable).
-    if (this.searchCount >= 8 && !this.toolLoopHintInjected) {
-      hints.push(`[Search storm] ${this.searchCount} search calls. STOP searching and synthesize.`);
-      this.toolLoopHintInjected = true;
-    }
-    if (this.searchCount >= 12) {
-      return { hints, stop: { note: 'Too many search requests. Synthesizing best answer.', contentLine: `\n\n[stuck] excessive web searching (${this.searchCount} calls) — stopping.\n` } };
-    }
-
-    return { hints };
-  }
-}
+/**
+ * Anti-loop guard for the agent reasoning loop.
+ *
+ * Extracted from chatStreamImpl (Phase 3). Holds the per-turn heuristic state
+ * (recent response texts, tool-call signatures, tool outcomes + once-only hint
+ * flags) and, after each round, returns a decision: zero or more system
+ * "hints" to nudge the model, and optionally a hard `stop` (assistant note +
+ * a user-visible content line). The agent loop applies the decision — the guard
+ * itself has no side effects, which makes every branch unit-testable.
+ *
+ * A fresh LoopGuard is created per turn (state must not leak across turns).
+ */
+
+import type { ToolCall } from '../llm';
+import {
+  toolCallSignature,
+  textSimilarity,
+  looksLikeFailedToolResult,
+  parseToolArgs,
+  SIG_WINDOW,
+  SIG_LOOP_HINT,
+  SIG_LOOP_HARDSTOP,
+} from '../agent_helpers';
+
+/** A hard stop: record `note` as an assistant message, then surface `contentLine`. */
+export interface GuardStop {
+  note: string;
+  contentLine: string;
+}
+
+/** The guard's decision for one round. `hints` apply in order; `stop` ends the turn. */
+export interface GuardDecision {
+  hints: string[];
+  stop?: GuardStop;
+}
+
+/** Minimal shape of a tool execution result the guard inspects. */
+export interface GuardExecResult {
+  toolName: string;
+  success: boolean;
+  result: string;
+}
+
+export class LoopGuard {
+  private recentResponseTexts: string[] = [];
+  private recentToolSigs: string[] = [];
+  private recentToolOutcomes: boolean[] = [];
+  private searchCount = 0; // cumulative search/fetch calls this turn (not window-bounded)
+  private repetitionHintInjected = false;
+  private toolLoopHintInjected = false; // shared by tool-signature loop + search-storm
+  private stuckHintInjected = false;
+
+  /**
+   * Observe one completed round. Mutates internal state and returns the
+   * hints/stop decision. Evaluation order (and the shared hint flag) mirrors
+   * the original inline logic exactly.
+   */
+  observe(
+    roundContent: string,
+    toolCallsReceived: ToolCall[],
+    execResults: Array<GuardExecResult | null>
+  ): GuardDecision {
+    const hints: string[] = [];
+
+    // 1. Narration-loop: response too similar to a recent one.
+    const normalizedRound = (roundContent || '').trim();
+    if (normalizedRound && this.recentResponseTexts.length > 0) {
+      const highSim = this.recentResponseTexts.slice(-2).some(prev => textSimilarity(normalizedRound, prev) >= 0.7);
+      if (highSim && !this.repetitionHintInjected) {
+        hints.push('[Stop narrating] Your last response is highly similar to your previous one. Stop writing prose. Either: (1) emit ONLY the next tool call, or (2) output the final deliverable.');
+        this.repetitionHintInjected = true;
+      }
+    }
+    this.recentResponseTexts.push(normalizedRound);
+    if (this.recentResponseTexts.length > 3) this.recentResponseTexts.shift();
+
+    // 2. Tool-signature loop: same call repeated within the window.
+    for (const tc of toolCallsReceived) {
+      const tName = tc.function.name;
+      if (['task_done', 'list_skills', 'use_skill'].includes(tName)) continue;
+      if (['web_search', 'fetch_page', 'http_get'].includes(tName)) this.searchCount++;
+      const rawArgs = tc.function.arguments;
+      const tArgs = typeof rawArgs === 'string' ? parseToolArgs(rawArgs) : rawArgs;
+      const sig = toolCallSignature(tName, tArgs);
+      if (sig) this.recentToolSigs.push(sig);
+    }
+    if (this.recentToolSigs.length > SIG_WINDOW) {
+      this.recentToolSigs.splice(0, this.recentToolSigs.length - SIG_WINDOW);
+    }
+    if (this.recentToolSigs.length > 0) {
+      const counts = new Map<string, number>();
+      for (const s of this.recentToolSigs) counts.set(s, (counts.get(s) || 0) + 1);
+      let topSig = '';
+      let topCount = 0;
+      for (const [s, c] of counts) { if (c > topCount) { topSig = s; topCount = c; } }
+      if (topCount >= SIG_LOOP_HINT && !this.toolLoopHintInjected) {
+        hints.push(`[Tool loop] You have called \`${topSig}\` ${topCount}x in the last ${this.recentToolSigs.length} tool calls — you are iterating without converging. STOP repeating it.`);
+        this.toolLoopHintInjected = true;
+      }
+      if (topCount >= SIG_LOOP_HARDSTOP) {
+        return { hints, stop: { note: `I have repeated \`${topSig}\` ${topCount} times without converging. Stopping.`, contentLine: `\n\n[stuck] tool \`${topSig}\` repeated ${topCount}x — stopping.` } };
+      }
+    }
+
+    // 3. Stuck-loop: most/all recent tool calls failed.
+    for (const r of execResults) {
+      if (!r || r.toolName === 'task_done') continue;
+      const failed = !r.success || (typeof r.result === 'string' && looksLikeFailedToolResult(r.result));
+      this.recentToolOutcomes.push(!failed);
+      // Keep 8 so the "all recent calls failed" (>=8) hard-stop below is reachable.
+      if (this.recentToolOutcomes.length > 8) this.recentToolOutcomes.shift();
+    }
+    if (!this.stuckHintInjected && this.recentToolOutcomes.length >= 5 &&
+        this.recentToolOutcomes.filter(Boolean).length <= 1) {
+      hints.push('[Recovery hint] Your last several tool calls have mostly failed. Synthesize a partial answer from what worked or ask the user for guidance.');
+      this.stuckHintInjected = true;
+    }
+    if (this.recentToolOutcomes.length >= 8 && this.recentToolOutcomes.every(x => !x)) {
+      return { hints, stop: { note: 'Every recent tool call failed. Please give me more context.', contentLine: '\n\n[stuck] every recent tool call failed — stopping.\n' } };
+    }
+
+    // 4. Search-storm: cumulative search/fetch calls this turn (not bounded by
+    // SIG_WINDOW, so the >=12 hard-stop is actually reachable).
+    if (this.searchCount >= 8 && !this.toolLoopHintInjected) {
+      hints.push(`[Search storm] ${this.searchCount} search calls. STOP searching and synthesize.`);
+      this.toolLoopHintInjected = true;
+    }
+    if (this.searchCount >= 12) {
+      return { hints, stop: { note: 'Too many search requests. Synthesizing best answer.', contentLine: `\n\n[stuck] excessive web searching (${this.searchCount} calls) — stopping.\n` } };
+    }
+
+    return { hints };
+  }
+}

package/src/core/agent/task.ts

@@ -1,100 +1,100 @@
-/**
- * Agent domain model — states and the Task DAG node.
- *
- * Extracted from the monolithic agent.ts (Phase 3). Pure, dependency-free,
- * and unit-testable in isolation. `agent.ts` re-exports these so external
- * importers of `../core/agent` are unaffected.
- */
-
-/** Lifecycle state of a running agent. */
-export enum AgentState {
-  IDLE = 'idle',
-  THINKING = 'thinking',
-  ACTING = 'acting',
-  WAITING = 'waiting',
-  ERROR = 'error',
-}
-
-/** Lifecycle state of an orchestrated task. */
-export enum TaskState {
-  PENDING = 'pending',
-  RUNNING = 'running',
-  COMPLETED = 'completed',
-  FAILED = 'failed',
-  SKIPPED = 'skipped',
-}
-
-/** Allowed task state transitions. A terminal state has no outgoing edges. */
-export const VALID_TRANSITIONS: Record<TaskState, Set<TaskState>> = {
-  [TaskState.PENDING]: new Set([TaskState.RUNNING, TaskState.SKIPPED, TaskState.FAILED]),
-  [TaskState.RUNNING]: new Set([TaskState.RUNNING, TaskState.COMPLETED, TaskState.FAILED]),
-  [TaskState.FAILED]: new Set([TaskState.RUNNING, TaskState.SKIPPED]),
-  [TaskState.COMPLETED]: new Set(),
-  [TaskState.SKIPPED]: new Set(),
-};
-
-/** A node in the orchestration DAG. */
-export class Task {
-  id: string;
-  description: string;
-  assignedTo: string | null = null;
-  parentId: string | null = null;
-  dependsOn: string[] = [];
-  status: TaskState = TaskState.PENDING;
-  priority: number = 0;
-  result: string | null = null;
-  metadata: Record<string, any> = {};
-
-  constructor(config: {
-    id: string;
-    description: string;
-    assignedTo?: string | null;
-    parentId?: string | null;
-    dependsOn?: string[];
-    status?: TaskState;
-    priority?: number;
-    result?: string | null;
-    metadata?: Record<string, any>;
-  }) {
-    this.id = config.id;
-    this.description = config.description;
-    this.assignedTo = config.assignedTo ?? null;
-    this.parentId = config.parentId ?? null;
-    this.dependsOn = config.dependsOn || [];
-    this.status = config.status ?? TaskState.PENDING;
-    this.priority = config.priority ?? 0;
-    this.result = config.result ?? null;
-    this.metadata = config.metadata || {};
-  }
-
-  transitionTo(newState: TaskState): void {
-    const allowed = VALID_TRANSITIONS[this.status] || new Set();
-    if (!allowed.has(newState)) {
-      throw new Error(
-        `Invalid task state transition: ${this.status} -> ${newState}`
-      );
-    }
-    this.status = newState;
-  }
-
-  get allDeps(): string[] {
-    const deps = [...this.dependsOn];
-    if (this.parentId && !deps.includes(this.parentId)) {
-      deps.push(this.parentId);
-    }
-    return deps;
-  }
-}
-
-/** Result of executing a single task. */
-export class TaskResult {
-  success: boolean;
-  content: string;
-  data: Record<string, any> = {};
-
-  constructor(success: boolean, content: string, data?: Record<string, any>) {
-    this.success = success;
-    this.content = content;
-    this.data = data || {};
-  }
-}
+/**
+ * Agent domain model — states and the Task DAG node.
+ *
+ * Extracted from the monolithic agent.ts (Phase 3). Pure, dependency-free,
+ * and unit-testable in isolation. `agent.ts` re-exports these so external
+ * importers of `../core/agent` are unaffected.
+ */
+
+/** Lifecycle state of a running agent. */
+export enum AgentState {
+  IDLE = 'idle',
+  THINKING = 'thinking',
+  ACTING = 'acting',
+  WAITING = 'waiting',
+  ERROR = 'error',
+}
+
+/** Lifecycle state of an orchestrated task. */
+export enum TaskState {
+  PENDING = 'pending',
+  RUNNING = 'running',
+  COMPLETED = 'completed',
+  FAILED = 'failed',
+  SKIPPED = 'skipped',
+}
+
+/** Allowed task state transitions. A terminal state has no outgoing edges. */
+export const VALID_TRANSITIONS: Record<TaskState, Set<TaskState>> = {
+  [TaskState.PENDING]: new Set([TaskState.RUNNING, TaskState.SKIPPED, TaskState.FAILED]),
+  [TaskState.RUNNING]: new Set([TaskState.RUNNING, TaskState.COMPLETED, TaskState.FAILED]),
+  [TaskState.FAILED]: new Set([TaskState.RUNNING, TaskState.SKIPPED]),
+  [TaskState.COMPLETED]: new Set(),
+  [TaskState.SKIPPED]: new Set(),
+};
+
+/** A node in the orchestration DAG. */
+export class Task {
+  id: string;
+  description: string;
+  assignedTo: string | null = null;
+  parentId: string | null = null;
+  dependsOn: string[] = [];
+  status: TaskState = TaskState.PENDING;
+  priority: number = 0;
+  result: string | null = null;
+  metadata: Record<string, any> = {};
+
+  constructor(config: {
+    id: string;
+    description: string;
+    assignedTo?: string | null;
+    parentId?: string | null;
+    dependsOn?: string[];
+    status?: TaskState;
+    priority?: number;
+    result?: string | null;
+    metadata?: Record<string, any>;
+  }) {
+    this.id = config.id;
+    this.description = config.description;
+    this.assignedTo = config.assignedTo ?? null;
+    this.parentId = config.parentId ?? null;
+    this.dependsOn = config.dependsOn || [];
+    this.status = config.status ?? TaskState.PENDING;
+    this.priority = config.priority ?? 0;
+    this.result = config.result ?? null;
+    this.metadata = config.metadata || {};
+  }
+
+  transitionTo(newState: TaskState): void {
+    const allowed = VALID_TRANSITIONS[this.status] || new Set();
+    if (!allowed.has(newState)) {
+      throw new Error(
+        `Invalid task state transition: ${this.status} -> ${newState}`
+      );
+    }
+    this.status = newState;
+  }
+
+  get allDeps(): string[] {
+    const deps = [...this.dependsOn];
+    if (this.parentId && !deps.includes(this.parentId)) {
+      deps.push(this.parentId);
+    }
+    return deps;
+  }
+}
+
+/** Result of executing a single task. */
+export class TaskResult {
+  success: boolean;
+  content: string;
+  data: Record<string, any> = {};
+
+  constructor(success: boolean, content: string, data?: Record<string, any>) {
+    this.success = success;
+    this.content = content;
+    this.data = data || {};
+  }
+}