clawed 2.3.7__tar.gz → 2.3.9__tar.gz

clawed-2.3.7/README.md → clawed-2.3.9/PKG-INFO

@@ -1,18 +1,136 @@
+Metadata-Version: 2.4
+Name: clawed
+Version: 2.3.9
+Summary: Free AI lesson planner for teachers. Learns your teaching voice from your curriculum files and generates complete lesson packages. Open source, runs locally, no subscription.
+Project-URL: Homepage, https://sirhanmacx.github.io/Claw-ED
+Project-URL: Documentation, https://github.com/SirhanMacx/Claw-ED#readme
+Project-URL: Repository, https://github.com/SirhanMacx/Claw-ED
+Project-URL: Bug Tracker, https://github.com/SirhanMacx/Claw-ED/issues
+Project-URL: Changelog, https://github.com/SirhanMacx/Claw-ED/blob/main/CHANGELOG.md
+Project-URL: PyPI, https://pypi.org/project/clawed/
+Author: Jon Maccarello & Claw-ED contributors
+License-Expression: MIT
+License-File: LICENSE
+Keywords: ai,ai-for-teachers,claw-ed,clawed,curriculum,edtech,education,hermes-agent,k12,lesson-plan-generator,lesson-planner,lesson-planning,lesson-plans,open-source-education,pedagogy,teacher-tools,teachers,teaching,teaching-assistant
+Classifier: Development Status :: 3 - Alpha
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Education
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Education
+Classifier: Topic :: Education :: Computer Aided Instruction (CAI)
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Requires-Dist: apscheduler<4.0,>=3.10.0
+Requires-Dist: fastapi<1.0,>=0.110.0
+Requires-Dist: httpx<1.0,>=0.25.0
+Requires-Dist: jinja2>=3.1.0
+Requires-Dist: json-repair>=0.30.0
+Requires-Dist: lxml>=4.9.0
+Requires-Dist: mcp>=1.0.0
+Requires-Dist: pydantic<3.0,>=2.0.0
+Requires-Dist: pymupdf>=1.23.0
+Requires-Dist: python-docx>=1.0.0
+Requires-Dist: python-multipart>=0.0.6
+Requires-Dist: python-pptx>=0.6.21
+Requires-Dist: pyyaml<7.0,>=6.0
+Requires-Dist: reportlab>=4.0.0
+Requires-Dist: rich>=13.0.0
+Requires-Dist: sse-starlette>=1.6.0
+Requires-Dist: typer<1.0,>=0.9.0
+Requires-Dist: uvicorn[standard]>=0.27.0
+Provides-Extra: all
+Requires-Dist: faster-whisper>=0.10.0; extra == 'all'
+Requires-Dist: keyring>=24.0.0; extra == 'all'
+Requires-Dist: onnxruntime>=1.16.0; extra == 'all'
+Requires-Dist: qrcode[pil]>=7.0; extra == 'all'
+Requires-Dist: textual>=0.56.0; extra == 'all'
+Requires-Dist: uvicorn[standard]>=0.27.0; extra == 'all'
+Provides-Extra: dev
+Requires-Dist: apscheduler<4.0,>=3.10.0; extra == 'dev'
+Requires-Dist: faster-whisper>=0.10.0; extra == 'dev'
+Requires-Dist: pytest-asyncio>=0.21.0; extra == 'dev'
+Requires-Dist: pytest-cov>=4.0; extra == 'dev'
+Requires-Dist: pytest>=7.0.0; extra == 'dev'
+Requires-Dist: ruff>=0.1.0; extra == 'dev'
+Provides-Extra: google
+Requires-Dist: google-api-python-client>=2.0.0; extra == 'google'
+Requires-Dist: google-auth-oauthlib>=1.0.0; extra == 'google'
+Provides-Extra: keyring
+Requires-Dist: keyring>=24.0.0; extra == 'keyring'
+Provides-Extra: memory
+Requires-Dist: onnxruntime>=1.16.0; extra == 'memory'
+Provides-Extra: pdf
+Requires-Dist: weasyprint>=60.0; extra == 'pdf'
+Provides-Extra: qr
+Requires-Dist: qrcode[pil]>=7.0; extra == 'qr'
+Provides-Extra: tui
+Requires-Dist: textual>=0.56.0; extra == 'tui'
+Provides-Extra: voice
+Requires-Dist: faster-whisper>=0.10.0; extra == 'voice'
+Description-Content-Type: text/markdown
+
 # Claw-ED
 
 Your personal AI teaching agent. Runs on your machine, learns your voice, works while you sleep.
 
-Built on the OpenClaw agent framework. Open source. MIT license.
+Built on the Hermes Agent framework (NousResearch). Open source. MIT license.
 
 <p align="center">
-  <a href="https://pypi.org/project/clawed/"><img src="https://img.shields.io/pypi/v/clawed?color=blue" alt="PyPI"></a>
-  <a href="https://python.org"><img src="https://img.shields.io/badge/Python-3.10+-blue" alt="Python 3.10+"></a>
+  <a href="https://pypi.org/project/clawed/"><img src="https://img.shields.io/pypi/v/clawed?color=blue" alt="PyPI version"></a>
+  <a href="https://pypi.org/project/clawed/"><img src="https://img.shields.io/pypi/pyversions/clawed" alt="Python versions"></a>
   <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-green" alt="MIT License"></a>
   <a href="https://github.com/SirhanMacx/Claw-ED/stargazers"><img src="https://img.shields.io/github/stars/SirhanMacx/Claw-ED" alt="GitHub stars"></a>
+  <a href="https://pepy.tech/project/clawed"><img src="https://static.pepy.tech/badge/clawed" alt="Downloads"></a>
 </p>
 
 ---
 
+## What's new in v2.3.9
+
+**Multi-agent lesson generation.** Instead of one LLM call doing everything, `--multi-agent` spawns a team of three specialized agents: a Researcher who finds primary sources and historical context, a Writer who drafts the lesson in your voice, and a Reviewer who scores quality on voice fidelity, pedagogy, and differentiation — sending it back for revision if any dimension scores below 7/10. The result: higher-quality lessons with better sources and stronger voice matching.
+
+```bash
+clawed lesson "The Missouri Compromise" -g 8 --multi-agent
+```
+
+**Claude Code integration.** Claw-ED now auto-reads OAuth tokens from Claude Code's credential store (`~/.claude/.credentials.json`). If you have Claude Code installed, Claw-ED just works — no separate API key setup needed, and tokens auto-refresh. Claw-ED also works as an MCP server that Claude Code can use directly:
+
+```bash
+clawed mcp-server  # Add to Claude Code as an MCP tool
+```
+
+**Continuous improvement pipeline.** The new `clawed train` command lets your AI fleet (or you) continuously improve lesson quality:
+
+```bash
+clawed train --drive          # Ingest new files from Google Drive, refine persona
+clawed train --path ./files   # Ingest local curriculum files
+clawed train --benchmark -n 5 # Generate 5 lessons and score them
+clawed train --full           # Drive ingest + benchmark
+```
+
+Persona refinement is now incremental — new voice markers merge with existing ones instead of overwriting. Your teaching identity gets richer over time.
+
+**OAuth auth fix.** All Anthropic API calls now correctly detect OAuth tokens and use Bearer authentication with Claude Code identity headers. Fixes the 401 errors that affected fleet agents and Claw-ED direct usage.
+
+---
+
+## What's new in v2.3.8
+
+**No more silent failures.** Every step of lesson generation now reports what happened — persona loading, material search, quality review, voice matching. If something fails, you'll know exactly what and why, with structured NLAH failure codes. Quality review runs automatically on every generated lesson and fails closed: if the review itself crashes, it reports failure instead of silently passing.
+
+**Stricter quality gates.** Lessons now require at least 6 guided notes, 3 exit ticket questions, and 2 primary sources with actual text. Topic drift is caught automatically. Voice match scoring compares generated lessons against your teaching persona.
+
+**Safer onboarding.** Teacher names and subjects are validated and truncated. Invalid grade levels get a re-prompt instead of being silently accepted. Demo mode can be forced with `CLAWED_DEMO=1` for presentations.
+
+**Async cleanup.** Background ingestion no longer risks crashing on Python 3.12+ from nested event loops. The fix is shared across all async call sites.
+
+---
+
 ## What's new in v2.3.7
 
 **Real images in every lesson.** Image specs are now required for every primary source and instruction section across all subjects. The LLM generates specific search queries ("Thomas Nast Boss Tweed political cartoon 1871") instead of leaving the field blank. Teacher images are found first using a three-stage progressive search (full query, individual keywords, subject fallback) with filename-weighted scoring across up to 150 candidates. External sources (Library of Congress, Wikimedia Commons, Unsplash) fill in the rest with subject-aware routing.
@@ -35,7 +153,7 @@ Built on the OpenClaw agent framework. Open source. MIT license.
 
 ## What is this?
 
-Claw-ED is a personal AI agent for teachers. It uses the same architecture as other OpenClaw agents -- SOUL.md for identity, layered memory, a workspace, pulse jobs, tool use, and autonomous operation. The difference is what it knows: your lesson plans, your teaching style, your standards, your students.
+Claw-ED is a personal AI agent for teachers. It uses the same architecture as Hermes Agent -- SOUL.md for identity, layered memory, a workspace, cron jobs, tool use, and autonomous operation. The difference is what it knows: your lesson plans, your teaching style, your standards, your students.
 
 You feed it your curriculum files and it learns your voice. Not a generic "teacher tone" -- your actual patterns, your vocabulary, the way you structure a Do Now or frame an essential question. Then it works: drafting lessons that sound like you wrote them, organizing your Google Drive, answering student questions in your voice, prepping your week before you wake up.
 

clawed-2.3.7/PKG-INFO → clawed-2.3.9/README.md

@@ -1,87 +1,60 @@
-Metadata-Version: 2.4
-Name: clawed
-Version: 2.3.7
-Summary: Claw-ED — personal AI teaching agent. Learns your voice, works while you sleep.
-Project-URL: Homepage, https://github.com/SirhanMacx/Claw-ED
-Project-URL: Documentation, https://github.com/SirhanMacx/Claw-ED#readme
-Project-URL: Issues, https://github.com/SirhanMacx/Claw-ED/issues
-Author: Jon Maccarello & Claw-ED contributors
-License-Expression: MIT
-License-File: LICENSE
-Keywords: ai,claw-ed,clawed,education,lesson-plans,teachers,teaching
-Classifier: Development Status :: 3 - Alpha
-Classifier: Intended Audience :: Education
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.10
-Classifier: Programming Language :: Python :: 3.11
-Classifier: Programming Language :: Python :: 3.12
-Classifier: Topic :: Education
-Requires-Python: >=3.10
-Requires-Dist: apscheduler<4.0,>=3.10.0
-Requires-Dist: fastapi<1.0,>=0.110.0
-Requires-Dist: httpx<1.0,>=0.25.0
-Requires-Dist: jinja2>=3.1.0
-Requires-Dist: json-repair>=0.30.0
-Requires-Dist: lxml>=4.9.0
-Requires-Dist: mcp>=1.0.0
-Requires-Dist: pydantic<3.0,>=2.0.0
-Requires-Dist: pymupdf>=1.23.0
-Requires-Dist: python-docx>=1.0.0
-Requires-Dist: python-multipart>=0.0.6
-Requires-Dist: python-pptx>=0.6.21
-Requires-Dist: pyyaml<7.0,>=6.0
-Requires-Dist: reportlab>=4.0.0
-Requires-Dist: rich>=13.0.0
-Requires-Dist: sse-starlette>=1.6.0
-Requires-Dist: typer<1.0,>=0.9.0
-Requires-Dist: uvicorn[standard]>=0.27.0
-Provides-Extra: all
-Requires-Dist: faster-whisper>=0.10.0; extra == 'all'
-Requires-Dist: keyring>=24.0.0; extra == 'all'
-Requires-Dist: onnxruntime>=1.16.0; extra == 'all'
-Requires-Dist: qrcode[pil]>=7.0; extra == 'all'
-Requires-Dist: textual>=0.56.0; extra == 'all'
-Requires-Dist: uvicorn[standard]>=0.27.0; extra == 'all'
-Provides-Extra: dev
-Requires-Dist: apscheduler<4.0,>=3.10.0; extra == 'dev'
-Requires-Dist: faster-whisper>=0.10.0; extra == 'dev'
-Requires-Dist: pytest-asyncio>=0.21.0; extra == 'dev'
-Requires-Dist: pytest-cov>=4.0; extra == 'dev'
-Requires-Dist: pytest>=7.0.0; extra == 'dev'
-Requires-Dist: ruff>=0.1.0; extra == 'dev'
-Provides-Extra: google
-Requires-Dist: google-api-python-client>=2.0.0; extra == 'google'
-Requires-Dist: google-auth-oauthlib>=1.0.0; extra == 'google'
-Provides-Extra: keyring
-Requires-Dist: keyring>=24.0.0; extra == 'keyring'
-Provides-Extra: memory
-Requires-Dist: onnxruntime>=1.16.0; extra == 'memory'
-Provides-Extra: pdf
-Requires-Dist: weasyprint>=60.0; extra == 'pdf'
-Provides-Extra: qr
-Requires-Dist: qrcode[pil]>=7.0; extra == 'qr'
-Provides-Extra: tui
-Requires-Dist: textual>=0.56.0; extra == 'tui'
-Provides-Extra: voice
-Requires-Dist: faster-whisper>=0.10.0; extra == 'voice'
-Description-Content-Type: text/markdown
-
 # Claw-ED
 
 Your personal AI teaching agent. Runs on your machine, learns your voice, works while you sleep.
 
-Built on the OpenClaw agent framework. Open source. MIT license.
+Built on the Hermes Agent framework (NousResearch). Open source. MIT license.
 
 <p align="center">
-  <a href="https://pypi.org/project/clawed/"><img src="https://img.shields.io/pypi/v/clawed?color=blue" alt="PyPI"></a>
-  <a href="https://python.org"><img src="https://img.shields.io/badge/Python-3.10+-blue" alt="Python 3.10+"></a>
+  <a href="https://pypi.org/project/clawed/"><img src="https://img.shields.io/pypi/v/clawed?color=blue" alt="PyPI version"></a>
+  <a href="https://pypi.org/project/clawed/"><img src="https://img.shields.io/pypi/pyversions/clawed" alt="Python versions"></a>
   <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-green" alt="MIT License"></a>
   <a href="https://github.com/SirhanMacx/Claw-ED/stargazers"><img src="https://img.shields.io/github/stars/SirhanMacx/Claw-ED" alt="GitHub stars"></a>
+  <a href="https://pepy.tech/project/clawed"><img src="https://static.pepy.tech/badge/clawed" alt="Downloads"></a>
 </p>
 
 ---
 
+## What's new in v2.3.9
+
+**Multi-agent lesson generation.** Instead of one LLM call doing everything, `--multi-agent` spawns a team of three specialized agents: a Researcher who finds primary sources and historical context, a Writer who drafts the lesson in your voice, and a Reviewer who scores quality on voice fidelity, pedagogy, and differentiation — sending it back for revision if any dimension scores below 7/10. The result: higher-quality lessons with better sources and stronger voice matching.
+
+```bash
+clawed lesson "The Missouri Compromise" -g 8 --multi-agent
+```
+
+**Claude Code integration.** Claw-ED now auto-reads OAuth tokens from Claude Code's credential store (`~/.claude/.credentials.json`). If you have Claude Code installed, Claw-ED just works — no separate API key setup needed, and tokens auto-refresh. Claw-ED also works as an MCP server that Claude Code can use directly:
+
+```bash
+clawed mcp-server  # Add to Claude Code as an MCP tool
+```
+
+**Continuous improvement pipeline.** The new `clawed train` command lets your AI fleet (or you) continuously improve lesson quality:
+
+```bash
+clawed train --drive          # Ingest new files from Google Drive, refine persona
+clawed train --path ./files   # Ingest local curriculum files
+clawed train --benchmark -n 5 # Generate 5 lessons and score them
+clawed train --full           # Drive ingest + benchmark
+```
+
+Persona refinement is now incremental — new voice markers merge with existing ones instead of overwriting. Your teaching identity gets richer over time.
+
+**OAuth auth fix.** All Anthropic API calls now correctly detect OAuth tokens and use Bearer authentication with Claude Code identity headers. Fixes the 401 errors that affected fleet agents and Claw-ED direct usage.
+
+---
+
+## What's new in v2.3.8
+
+**No more silent failures.** Every step of lesson generation now reports what happened — persona loading, material search, quality review, voice matching. If something fails, you'll know exactly what and why, with structured NLAH failure codes. Quality review runs automatically on every generated lesson and fails closed: if the review itself crashes, it reports failure instead of silently passing.
+
+**Stricter quality gates.** Lessons now require at least 6 guided notes, 3 exit ticket questions, and 2 primary sources with actual text. Topic drift is caught automatically. Voice match scoring compares generated lessons against your teaching persona.
+
+**Safer onboarding.** Teacher names and subjects are validated and truncated. Invalid grade levels get a re-prompt instead of being silently accepted. Demo mode can be forced with `CLAWED_DEMO=1` for presentations.
+
+**Async cleanup.** Background ingestion no longer risks crashing on Python 3.12+ from nested event loops. The fix is shared across all async call sites.
+
+---
+
 ## What's new in v2.3.7
 
 **Real images in every lesson.** Image specs are now required for every primary source and instruction section across all subjects. The LLM generates specific search queries ("Thomas Nast Boss Tweed political cartoon 1871") instead of leaving the field blank. Teacher images are found first using a three-stage progressive search (full query, individual keywords, subject fallback) with filename-weighted scoring across up to 150 candidates. External sources (Library of Congress, Wikimedia Commons, Unsplash) fill in the rest with subject-aware routing.
@@ -104,7 +77,7 @@ Built on the OpenClaw agent framework. Open source. MIT license.
 
 ## What is this?
 
-Claw-ED is a personal AI agent for teachers. It uses the same architecture as other OpenClaw agents -- SOUL.md for identity, layered memory, a workspace, pulse jobs, tool use, and autonomous operation. The difference is what it knows: your lesson plans, your teaching style, your standards, your students.
+Claw-ED is a personal AI agent for teachers. It uses the same architecture as Hermes Agent -- SOUL.md for identity, layered memory, a workspace, cron jobs, tool use, and autonomous operation. The difference is what it knows: your lesson plans, your teaching style, your standards, your students.
 
 You feed it your curriculum files and it learns your voice. Not a generic "teacher tone" -- your actual patterns, your vocabulary, the way you structure a Do Now or frame an essential question. Then it works: drafting lessons that sound like you wrote them, organizing your Google Drive, answering student questions in your voice, prepping your week before you wake up.
 

{clawed-2.3.7 → clawed-2.3.9}/clawed/__init__.py

@@ -17,7 +17,7 @@ if hasattr(sys.stderr, "reconfigure"):
     except Exception:
         pass
 
-__version__ = "2.3.7"
+__version__ = "2.3.9"
 __author__ = "Jon Maccarello & Claw-ED contributors"
 __description__ = "Personal AI teaching agent. Learns your voice, works while you sleep."
 

{clawed-2.3.7 → clawed-2.3.9}/clawed/_legacy_gateway.py

@@ -350,7 +350,7 @@ class Gateway:
 
     async def _status_response(self, teacher_id: str) -> GatewayResponse:
         try:
-            from clawed.openclaw_plugin import _show_status
+            from clawed.hermes_plugin import _show_status
             from clawed.state import TeacherSession
             session = TeacherSession.load(teacher_id)
             return GatewayResponse(text=_show_status(session))

{clawed-2.3.7 → clawed-2.3.9}/clawed/agent.py

@@ -13,6 +13,29 @@ from __future__ import annotations
 
 import json
 import logging
+
+
+def _anthropic_headers(api_key: str) -> dict[str, str]:
+    """Build Anthropic API headers, auto-detecting OAuth vs regular API keys.
+
+    OAuth tokens (sk-ant-oat01-*) need Bearer auth + Claude Code identity
+    headers. Regular API keys (sk-ant-api*) use x-api-key.
+    """
+    is_oauth = api_key.startswith("sk-ant-") and not api_key.startswith("sk-ant-api")
+    if is_oauth:
+        return {
+            "authorization": f"Bearer {api_key}",
+            "anthropic-version": "2023-06-01",
+            "anthropic-beta": "interleaved-thinking-2025-05-14,claude-code-20250219,oauth-2025-04-20",
+            "user-agent": "claude-cli/1.0.0 (external, cli)",
+            "x-app": "cli",
+            "content-type": "application/json",
+        }
+    return {
+        "x-api-key": api_key,
+        "anthropic-version": "2023-06-01",
+        "content-type": "application/json",
+    }
 from typing import Any
 
 from clawed.models import AppConfig, LLMProvider
@@ -137,11 +160,7 @@ async def _anthropic_with_tools(
     async with httpx.AsyncClient(timeout=120.0) as client:
         resp = await client.post(
             "https://api.anthropic.com/v1/messages",
-            headers={
-                "x-api-key": api_key,
-                "anthropic-version": "2023-06-01",
-                "content-type": "application/json",
-            },
+            headers=_anthropic_headers(api_key),
             json={
                 "model": config.anthropic_model,
                 "max_tokens": 4096,

{clawed-2.3.7 → clawed-2.3.9}/clawed/agent_core/core.py

@@ -339,7 +339,9 @@ class Gateway:
         # 2a. Load SOUL.md if available
         soul_context = ""
         try:
-            soul_path = Path.home() / ".eduagent" / "workspace" / "SOUL.md"
+            import os
+            data_dir = os.environ.get("EDUAGENT_DATA_DIR", str(Path.home() / ".eduagent"))
+            soul_path = Path(data_dir) / "workspace" / "SOUL.md"
             if soul_path.exists():
                 soul_context = soul_path.read_text(encoding="utf-8")[:2000]
         except Exception:

{clawed-2.3.7 → clawed-2.3.9}/clawed/agent_core/tools/generate_lesson_bundle.py

@@ -6,6 +6,7 @@ from pathlib import Path
 from typing import Any
 
 from clawed.agent_core.context import AgentContext, ToolResult
+from clawed.failure_codes import FailureCode
 
 logger = logging.getLogger(__name__)
 
@@ -89,14 +90,18 @@ class GenerateLessonBundleTool:
             f"this usually takes 2-4 minutes. I'll send everything when it's ready!"
         )
 
+        from clawed.generation_report import GenerationReport
+        report = GenerationReport()
+
         # ── Load config & persona from context ───────────────────────
         config = context.config
         persona = TeacherPersona()
         if context.persona:
             try:
                 persona = TeacherPersona(**context.persona)
-            except Exception:
-                pass
+            except Exception as e:
+                logger.warning("NLAH_FAILURE=%s: %s", FailureCode.PERSONA_PARSE_ERROR, e)
+                report.warnings.append(f"[{FailureCode.PERSONA_PARSE_ERROR}] Could not parse persona: {e}")
 
         # ── Load state standards if teacher profile has a state ───────
         state = ""
@@ -127,7 +132,8 @@ class GenerateLessonBundleTool:
                     len(assets), len(yt_links), topic,
                 )
         except Exception as e:
-            logger.debug("Asset search failed: %s", e)
+            logger.warning("NLAH_FAILURE=%s: %s", FailureCode.ASSET_SEARCH_FAILED, e)
+            report.warnings.append(f"[{FailureCode.ASSET_SEARCH_FAILED}] Asset search failed: {e}")
 
         # KB chunk-level search (text excerpts)
         try:
@@ -162,7 +168,8 @@ class GenerateLessonBundleTool:
                         )
                     logger.info("KB search found %d relevant chunks for '%s'", len(kb_parts), topic)
         except Exception as e:
-            logger.debug("KB search failed: %s", e)
+            logger.warning("NLAH_FAILURE=%s: %s", FailureCode.KB_SEARCH_FAILED, e)
+            report.warnings.append(f"[{FailureCode.KB_SEARCH_FAILED}] KB search failed: {e}")
 
         # ── Build a UnitPlan with standards ──────────────────────────
         description = f"Introduction to {topic}"
@@ -205,14 +212,12 @@ class GenerateLessonBundleTool:
                 teacher_materials=kb_prompt_section,
             )
         except Exception as e:
-            return ToolResult(text=f"Failed to generate lesson: {e}")
+            logger.error("NLAH_FAILURE=%s: %s", FailureCode.API_FAILURE, e)
+            return ToolResult(text=f"[{FailureCode.API_FAILURE}] Failed to generate lesson: {type(e).__name__}")
 
         # ── Validate ──────────────────────────────────────────────────
-        from clawed.generation_report import GenerationReport
         from clawed.validation import check_self_contained, validate_alignment, validate_master_content
 
-        report = GenerationReport()
-
         mc_errors = validate_master_content(master, topic)
         for err in mc_errors:
             report.warnings.append(err)
@@ -283,6 +288,45 @@ class GenerateLessonBundleTool:
             logger.error("Slides compile failed: %s", e)
             errors.append(f"Slideshow PPTX failed: {e}")
 
+        # ── Quality review (NLAH Stage 4 — non-blocking) ────────────
+        try:
+            from clawed.llm import LLMClient
+            from clawed.quality import score_voice_match
+
+            llm = LLMClient(config=config)
+            master_json = master.model_dump_json(indent=2)[:3000]
+            review = await llm.review_lesson_package(
+                lesson_json=master_json,
+                standards_present=bool(standards_list),
+                has_handout=len(generated_files) >= 2,
+                has_slideshow=len(generated_files) >= 3,
+            )
+            report.quality_review_passed = review.get("passed", False)
+            report.quality_review_issues = review.get("issues", [])
+            if not report.quality_review_passed:
+                for issue in report.quality_review_issues:
+                    report.warnings.append(f"[REVIEW] {issue}")
+                logger.info("Quality review: FAILED — %d issues", len(report.quality_review_issues))
+            else:
+                logger.info("Quality review: PASSED")
+
+            # Voice match scoring
+            persona_ctx = persona.to_prompt_context() if persona else ""
+            all_text = " ".join(s.content for s in master.direct_instruction)
+            voice_score = await score_voice_match(all_text, persona_ctx, llm)
+            report.voice_check_passed = voice_score >= 3.0
+            if voice_score < 3.0:
+                report.warnings.append(
+                    f"[{FailureCode.VOICE_MISMATCH}] Voice match score: {voice_score:.1f}/5.0"
+                )
+                logger.warning("Voice match: %.1f/5.0 — below threshold", voice_score)
+            else:
+                logger.info("Voice match: %.1f/5.0", voice_score)
+        except Exception as e:
+            logger.warning("Quality review/voice check failed: %s", e)
+            report.quality_review_passed = False
+            report.quality_review_issues = [f"Review failed: {type(e).__name__}"]
+
         # ── Build response ─────────────────────────────────────────────
         lines = []
 

{clawed-2.3.7 → clawed-2.3.9}/clawed/api/server.py

@@ -266,7 +266,9 @@ def create_app() -> FastAPI:
         filter_html = ""
         if all_subjects:
             opts = "".join(
-                f"<option value='{html_mod.escape(s)}' {'selected' if s == subject_filter else ''}>{html_mod.escape(s)}</option>"
+                f"<option value='{html_mod.escape(s)}'"
+                f" {'selected' if s == subject_filter else ''}"
+                f">{html_mod.escape(s)}</option>"
                 for s in all_subjects
             )
             filter_html += (
@@ -275,7 +277,9 @@ def create_app() -> FastAPI:
             )
         if all_grades:
             opts = "".join(
-                f"<option value='{html_mod.escape(g)}' {'selected' if g == grade_filter else ''}>{html_mod.escape(g)}</option>"
+                f"<option value='{html_mod.escape(g)}'"
+                f" {'selected' if g == grade_filter else ''}"
+                f">{html_mod.escape(g)}</option>"
                 for g in all_grades
             )
             filter_html += (

clawed-2.3.9/clawed/async_utils.py

@@ -0,0 +1,22 @@
+"""Shared async utility for running coroutines from sync or async contexts."""
+from __future__ import annotations
+
+import asyncio
+
+
+def run_async_safe(coro):
+    """Run an async coroutine, handling both sync and async calling contexts.
+
+    When called from inside a running event loop (e.g., agent_core tools),
+    uses a thread to avoid nested asyncio.run() errors.  When called from
+    plain sync code (CLI), uses asyncio.run() directly.
+    """
+    try:
+        asyncio.get_running_loop()
+        # We're inside an event loop — run in a worker thread
+        import concurrent.futures
+        with concurrent.futures.ThreadPoolExecutor(max_workers=1) as pool:
+            return pool.submit(asyncio.run, coro).result(timeout=30)
+    except RuntimeError:
+        # No running loop — safe to use asyncio.run()
+        return asyncio.run(coro)

{clawed-2.3.7 → clawed-2.3.9}/clawed/cli.py

@@ -39,6 +39,7 @@ from clawed.commands.generate import generate_app
 from clawed.commands.queue import queue_app
 from clawed.commands.schedule_cmd import schedule_app
 from clawed.commands.sub import sub_app
+from clawed.commands.train import train_app
 from clawed.commands.workspace_cmd import workspace_app
 
 # ── Build the main app ──────────────────────────────────────────────────
@@ -241,6 +242,7 @@ app.add_typer(class_app, name="class")
 app.add_typer(queue_app, name="queue")
 app.add_typer(workspace_app, name="workspace")
 app.add_typer(schedule_app, name="schedule")
+app.add_typer(train_app, name="train")
 
 # ── Register top-level commands from sub-modules ────────────────────────
 # Commands from generate_app, export_app, and bot_app are registered as

{clawed-2.3.7 → clawed-2.3.9}/clawed/commands/generate.py

@@ -298,6 +298,10 @@ def lesson(
     homework: bool = typer.Option(
         True, "--homework/--no-homework", help="Include homework"
     ),
+    multi_agent: bool = typer.Option(
+        False, "--multi-agent/--no-multi-agent",
+        help="Use multi-agent pipeline (researcher→writer→reviewer) for higher quality",
+    ),
     fmt: str = typer.Option("markdown", "--format", "-f", help="Export format: markdown, pptx, docx, pdf, handout"),
 ):
     """Generate a detailed daily lesson plan.
@@ -306,6 +310,9 @@ def lesson(
     Standalone mode (no unit plan needed):
         clawed lesson "The American Revolution" --grade 8 --subject "social studies"
 
+    Multi-agent mode (higher quality, 3x slower):
+        clawed lesson "The Missouri Compromise" -g 8 --multi-agent
+
     From a unit plan:
         clawed lesson "Photosynthesis" --unit-file output/unit_photosynthesis.json -n 1
     """
@@ -388,18 +395,46 @@ def lesson(
 
     with _safe_progress(console=console) as progress:
         task = progress.add_task(
-            f"Generating lesson {lesson_num}...", total=None
+            f"Generating lesson {lesson_num}{'  [multi-agent]' if multi_agent else ''}...",
+            total=None,
         )
         try:
-            daily = _run_async(
-                generate_lesson(
-                    lesson_number=lesson_num,
-                    unit=unit_plan,
-                    persona=persona,
-                    include_homework=homework,
-                    teacher_materials=kb_prompt_section,
+            if multi_agent:
+                # Multi-agent pipeline: researcher → writer → reviewer
+                from clawed.multi_agent import multi_agent_generate_master_content
+                from clawed.compile_teacher import compile_teacher_view
+                console.print("[dim]Using multi-agent pipeline (researcher→writer→reviewer)...[/dim]")
+                master = _run_async(
+                    multi_agent_generate_master_content(
+                        topic=topic,
+                        grade=grade,
+                        subject=subject,
+                        persona=persona,
+                        config=None,
+                        unit_context=kb_prompt_section,
+                    )
+                )
+                if master is None:
+                    console.print("[yellow]Multi-agent returned None, falling back to single-agent...[/yellow]")
+                    daily = _run_async(
+                        generate_lesson(
+                            lesson_number=lesson_num, unit=unit_plan,
+                            persona=persona, include_homework=homework,
+                            teacher_materials=kb_prompt_section,
+                        )
+                    )
+                else:
+                    daily = compile_teacher_view(master)
+            else:
+                daily = _run_async(
+                    generate_lesson(
+                        lesson_number=lesson_num,
+                        unit=unit_plan,
+                        persona=persona,
+                        include_homework=homework,
+                        teacher_materials=kb_prompt_section,
+                    )
                 )
-            )
         except (RuntimeError, ValueError) as e:
             console.print(f"[red]Generation failed:[/red] {e}")
             console.print("[dim]Run with --debug for full details[/dim]")