wogiflow 2.30.4 → 2.31.1

package/scripts/hooks/entry/claude-code/user-prompt-submit.js

@@ -1,289 +1,21 @@
 #!/usr/bin/env node
 
 /**
- * Wogi Flow - Claude Code UserPromptSubmit Hook
+ * Wogi Flow - Claude Code UserPromptSubmit Hook (thin entry)
  *
- * Called when user submits a prompt (before processing).
- * Enforces implementation gate - blocks implementation requests without active task.
+ * All UserPromptSubmit business logic lives in
+ * scripts/hooks/core/user-prompt-orchestrator.js. This entry dispatches.
+ *
+ * Per .claude/rules/architecture/hook-three-layer.md: entry files ≤ 120 LOC,
+ * ≤ 2 core/ imports, no inline business logic. wf-6e31850e A-3 extracted
+ * the prior 293-LOC body into core/user-prompt-orchestrator.js.
  */
 
-const fs = require('node:fs');
-const { checkImplementationGate } = require('../../core/implementation-gate');
-const { checkResearchRequirement } = require('../../core/research-gate');
-const { setRoutingPending, clearRoutingPending, ROUTING_CLEARED_PATH } = require('../../core/routing-gate');
-const { getPhaseContextPrompt } = require('../../core/phase-gate');
-const { buildOverdueContext } = require('../../core/overdue-dispatches');
-const { getDossierInjection } = require('../../core/feature-dossier-gate');
-const {
-  shouldForceExtractReview,
-  buildEnforcementMessage,
-  markLongInputPending
-} = require('../../core/long-input-enforcement');
-const { markSkillPending, loadDurableSession } = require('../../../flow-durable-session');
-const { captureCurrentPrompt } = require('../../../flow-prompt-capture');
-const { spawnBackgroundDetection } = require('../../../flow-correction-detector');
-const { getConfig } = require('../../../flow-utils');
+const { orchestrateUserPromptSubmit } = require('../../core/user-prompt-orchestrator');
 const { runHook } = require('../shared/hook-runner');
 
 runHook('UserPromptSubmit', async ({ input, parsedInput }) => {
-  // Handle empty input gracefully
-  if (!input || Object.keys(input).length === 0) {
-    return { __raw: true, continue: true, hookSpecificOutput: { hookEventName: 'UserPromptSubmit' } };
-  }
-
-  const prompt = parsedInput.prompt;
-  const source = parsedInput.source;
-
-  // wf-729ab5c0 follow-up — clear pending-question marker on user response.
-  // This unblocks a deferred task-boundary restart. The AI may have asked a
-  // question via `flow ask "..."` after task completion; user's response
-  // releases the deferral, and the next Stop hook will fire the restart.
-  try {
-    const { clearPendingQuestion } = require('../../../flow-ask');
-    const r = clearPendingQuestion();
-    if (r.wasPresent && process.env.DEBUG) {
-      console.error(`[UserPromptSubmit] Cleared pending-question marker — restart deferral released`);
-    }
-  } catch (_err) { /* non-fatal */ }
-
-  // v4.1: Detect skill commands that need execution tracking
-  if (typeof prompt === 'string') {
-    const skillMatch = prompt.match(/^\/(wogi-bulk|wogi-start)\b/i);
-    if (skillMatch) {
-      const skillName = skillMatch[1].toLowerCase();
-      markSkillPending(skillName, { prompt });
-      if (process.env.DEBUG) {
-        console.error(`[Hook] Marked /${skillName} as pending execution`);
-      }
-    }
-  }
-
-  // Load config once for feature flag checks
-  let hookConfig;
-  try {
-    hookConfig = getConfig();
-  } catch (_err) {
-    hookConfig = {};
-  }
-
-  // v5.0: Capture prompt for learning system (non-blocking)
-  if (hookConfig.hooks?.rules?.intelligence?.promptCapture?.enabled !== false) {
-    if (typeof prompt === 'string' && prompt.trim().length > 0) {
-      setImmediate(() => {
-        try {
-          captureCurrentPrompt(prompt);
-        } catch (err) {
-          if (process.env.DEBUG) {
-            console.error(`[Hook] Prompt capture failed: ${err.message}`);
-          }
-        }
-      });
-    }
-  }
-
-  // wf-5cd71b1f: Research-required classifier — detect diagnostic prompts
-  // (Tier 2/3 from CLAUDE.md routing) that require evidence-reading before
-  // the AI answers. Writes a turn-scoped marker that the Stop hook checks;
-  // if the AI answered with text-only and no Read calls against evidence
-  // paths, the Stop hook re-prompts forcing a redo. Fail-open throughout.
-  if (typeof prompt === 'string' && prompt.trim().length > 0) {
-    try {
-      const { applyClassification: applyResearchClassification } = require('../../core/research-required-classifier');
-      const r = applyResearchClassification(prompt, hookConfig);
-      if (r.applied && process.env.DEBUG) {
-        console.error(`[Hook] Research-required classifier: category=${r.category}, match="${r.match}"`);
-      }
-    } catch (err) {
-      if (process.env.DEBUG) {
-        console.error(`[Hook] Research-required classifier failed: ${err.message}`);
-      }
-    }
-  }
-
-  // wf-b8839d99 (replaces wf-f9912af6 regex classifier): AI-based deferral-
-  // intent classifier. Calls Haiku to interpret the user's prompt. NEGATIVE
-  // ("fix all", "I don't like tech debt", any phrasing) writes a no-defer-pin;
-  // POSITIVE ("defer F5", "option 2", "ship as-is") writes a scoped auth
-  // marker. The marker now captures the verbatim user excerpt SEPARATELY from
-  // the AI's interpretation — ending the false-attribution failure shape.
-  // Fail-open throughout: classifier errors / missing API key → no state
-  // change (status quo holds; gate's default-restrictive behavior preserved).
-  if (typeof prompt === 'string' && prompt.trim().length > 0) {
-    try {
-      const { applyClassification } = require('../../core/deferral-classifier');
-      const r = await applyClassification(prompt, hookConfig);
-      if (r.applied && process.env.DEBUG) {
-        console.error(`[Hook] Deferral classifier (AI): intent=${r.intent}, confidence=${r.confidence}, standing=${r.standing}, scope=${JSON.stringify(r.scope)}`);
-      } else if (process.env.DEBUG && r.reason) {
-        console.error(`[Hook] Deferral classifier (AI): no-op — ${r.reason}`);
-      }
-    } catch (err) {
-      if (process.env.DEBUG) {
-        console.error(`[Hook] Deferral classifier failed: ${err.message}`);
-      }
-    }
-  }
-
-  // v5.1->v7.0: Detect corrections for learning system (AI-only, non-blocking)
-  if (hookConfig.hooks?.rules?.intelligence?.correctionDetection?.enabled !== false) {
-    if (typeof prompt === 'string' && prompt.trim().length > 0) {
-      try {
-        const session = loadDurableSession();
-        spawnBackgroundDetection(prompt, session?.taskId || '');
-      } catch (err) {
-        if (process.env.DEBUG) {
-          console.error(`[Hook] Correction detection spawn failed: ${err.message}`);
-        }
-      }
-    }
-  }
-
-  // v6.0: Set routing-pending flag for routing gate enforcement
-  const isWogiCommand = typeof prompt === 'string' && /^\/wogi-[a-z0-9-]+\b/i.test(prompt.trim());
-  if (!isWogiCommand) {
-    // v8.1: Delete any stale cleared marker from previous turns.
-    try {
-      fs.unlinkSync(ROUTING_CLEARED_PATH);
-    } catch (err) {
-      if (err.code !== 'ENOENT' && process.env.DEBUG) {
-        console.error(`[Hook] Failed to delete cleared marker: ${err.message}`);
-      }
-    }
-
-    try {
-      setRoutingPending();
-    } catch (err) {
-      if (process.env.DEBUG) {
-        console.error(`[Hook] Routing gate set failed: ${err.message}`);
-      }
-    }
-  } else {
-    // v6.2: Actively CLEAR any existing routing flag when user explicitly types a /wogi-* command.
-    try {
-      clearRoutingPending();
-      if (process.env.DEBUG) {
-        console.error(`[Hook] Cleared routing flag — prompt is a /wogi-* command`);
-      }
-    } catch (err) {
-      if (process.env.DEBUG) {
-        console.error(`[Hook] Routing gate clear failed: ${err.message}`);
-      }
-    }
-  }
-
-  // Phase context injection
-  let phasePrompt = null;
-  try {
-    const phaseContext = getPhaseContextPrompt();
-    if (phaseContext.inject && phaseContext.prompt) {
-      phasePrompt = phaseContext.prompt;
-    }
-  } catch (err) {
-    if (process.env.DEBUG) {
-      console.error(`[Hook] Phase context injection failed: ${err.message}`);
-    }
-  }
-
-  // wf-557cf08a — Feature dossier + logic rules auto-injection.
-  // Surfaces canonical per-feature knowledge and cross-cutting logic rules
-  // into the phase prompt so Claude doesn't have to fetch them under token
-  // pressure. Fail-open: returns null on any error.
-  let dossierPrompt = null;
-  try {
-    dossierPrompt = getDossierInjection();
-  } catch (err) {
-    if (process.env.DEBUG) {
-      console.error(`[Hook] Dossier injection failed: ${err.message}`);
-    }
-  }
-  if (dossierPrompt) {
-    phasePrompt = phasePrompt ? `${phasePrompt}\n\n${dossierPrompt}` : dossierPrompt;
-  }
-
-  // Check research gate first (before implementation gate)
-  const researchResult = checkResearchRequirement({
-    prompt,
-    source
-  });
-
-  // Check implementation gate
-  let coreResult = checkImplementationGate({
-    prompt,
-    source
-  });
-
-  // If research protocol should be injected, add it to system reminder
-  if (researchResult.injectProtocol && researchResult.protocolSteps) {
-    coreResult = {
-      ...coreResult,
-      systemReminder: researchResult.protocolSteps,
-      researchTriggered: true,
-      questionType: researchResult.questionType,
-      suggestedDepth: researchResult.suggestedDepth
-    };
-  } else if (researchResult.warning && coreResult.allowed) {
-    coreResult = {
-      ...coreResult,
-      warning: true,
-      researchWarning: researchResult.message,
-      suggestedCommand: researchResult.suggestedCommand
-    };
-  }
-
-  // Inject phase-specific context prompt
-  if (phasePrompt) {
-    coreResult = {
-      ...coreResult,
-      phasePrompt
-    };
-  }
-
-  // wf-d3e67abe — surface overdue workspace dispatches (silent worker deaths)
-  // to the manager model before it processes the next prompt. Manager-only;
-  // fail-open (buildOverdueContext returns null on any error or wrong scope).
-  try {
-    const overduePrompt = buildOverdueContext();
-    if (overduePrompt) {
-      coreResult = {
-        ...coreResult,
-        overduePrompt
-      };
-    }
-  } catch (err) {
-    if (process.env.DEBUG) {
-      console.error(`[Hook] Overdue dispatches check failed: ${err.message}`);
-    }
-  }
-
-  // P11.5 mechanical enforcement (2026-04-27): long-form prompts without
-  // source-link are forced through /wogi-extract-review. This is the
-  // mechanical layer that complements the methodology rule. Applies in
-  // worker mode (channel-dispatch with no source-link — wogi-hub failure
-  // shape) AND in any session that receives a long task-creating prompt
-  // without preserved source.
-  try {
-    const enforce = shouldForceExtractReview({ text: prompt, source });
-    if (enforce.forced) {
-      const msg = buildEnforcementMessage(enforce.reason, enforce.level);
-      coreResult = {
-        ...coreResult,
-        longInputEnforcement: msg
-      };
-      markLongInputPending({
-        level: enforce.level,
-        reason: enforce.reason,
-        promptPreview: typeof prompt === 'string' ? prompt.slice(0, 200) : '(non-string)',
-        source: source || null,
-        repoName: process.env.WOGI_REPO_NAME || null
-      });
-    }
-  } catch (err) {
-    if (process.env.DEBUG) {
-      console.error(`[Hook] Long-input enforcement check failed: ${err.message}`);
-    }
-  }
-
-  return coreResult;
+  return await orchestrateUserPromptSubmit({ input, parsedInput });
 }, {
   failMode: 'block',
   failOutput: {