aigroup-workflow 2.1.2 → 2.2.0

package/package.json

@@ -1,39 +1,39 @@
-{
-  "name": "aigroup-workflow",
-  "version": "2.1.2",
-  "description": "AI 团队协作框架 — 通过角色派遣、工作流管道和 Harness 传感器驱动 AI 协作开发",
-  "type": "module",
-  "bin": {
-    "aigroup": "bin/aigroup.mjs",
-    "aig": "bin/aigroup.mjs"
-  },
-  "files": [
-    "bin/",
-    "cli/",
-    "docs/",
-    "agents/",
-    "skills/",
-    "scripts/",
-    "manifests/",
-    ".claude/commands/",
-    ".claude/hooks.json",
-    ".claude-plugin/",
-    ".codex/",
-    ".codex-plugin/",
-    "CLAUDE.md",
-    "AGENTS.md"
-  ],
-  "keywords": [
-    "claude-code",
-    "ai-agent",
-    "workflow",
-    "harness",
-    "cli"
-  ],
-  "author": "shuaishuai-li",
-  "license": "MIT",
-  "engines": {
-    "node": ">=18.0.0"
-  },
-  "dependencies": {}
-}
+{
+  "name": "aigroup-workflow",
+  "version": "2.2.0",
+  "description": "AI 团队协作框架 — 通过角色派遣、工作流管道和 Harness 传感器驱动 AI 协作开发",
+  "type": "module",
+  "bin": {
+    "aigroup": "bin/aigroup.mjs",
+    "aig": "bin/aigroup.mjs"
+  },
+  "files": [
+    "bin/",
+    "cli/",
+    "docs/",
+    "agents/",
+    "skills/",
+    "scripts/",
+    "manifests/",
+    ".claude/commands/",
+    ".claude/hooks.json",
+    ".claude-plugin/",
+    ".codex/",
+    ".codex-plugin/",
+    "CLAUDE.md",
+    "AGENTS.md"
+  ],
+  "keywords": [
+    "claude-code",
+    "ai-agent",
+    "workflow",
+    "harness",
+    "cli"
+  ],
+  "author": "shuaishuai-li",
+  "license": "MIT",
+  "engines": {
+    "node": ">=18.0.0"
+  },
+  "dependencies": {}
+}

package/scripts/orchestration/lib/orchestrator.cjs

@@ -164,6 +164,39 @@ function appendHandoff(sessionName, workerName, sectionTitle, content) {
   return artifacts.handoff;
 }
 
+const STANDARD_SECTIONS = ['Summary', 'Files Changed', 'Validation', 'Follow-ups'];
+
+function replaceStandardSection(body, title, content) {
+  const trimmed = (content || '').trim() || '- _未提供_';
+  const escaped = title.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+  const re = new RegExp(`(^|\\n)## ${escaped}\\n[\\s\\S]*?(?=\\n## |$)`);
+  const replacement = `$1## ${title}\n${trimmed}\n`;
+  if (re.test(body)) return body.replace(re, replacement);
+  return body.replace(/\s+$/, '') + `\n\n## ${title}\n${trimmed}\n`;
+}
+
+function completeHandoff(sessionName, workerName, fields = {}) {
+  const artifacts = workerArtifacts(sessionName, workerName);
+  let body = fs.existsSync(artifacts.handoff)
+    ? fs.readFileSync(artifacts.handoff, 'utf8')
+    : buildHandoffFile({ workerName });
+  const map = {
+    Summary: fields.summary,
+    'Files Changed': fields.files,
+    Validation: fields.validation,
+    'Follow-ups': fields.followUps
+  };
+  for (const title of STANDARD_SECTIONS) {
+    if (map[title] !== undefined && map[title] !== null) {
+      body = replaceStandardSection(body, title, map[title]);
+    }
+  }
+  const stamp = `\n<!-- finalized: ${timestamp()} -->\n`;
+  body = body.replace(/\n<!-- finalized: [^>]+ -->\n?/g, '');
+  writeFile(artifacts.handoff, body.replace(/\s+$/, '') + stamp);
+  return artifacts.handoff;
+}
+
 module.exports = {
   COORDINATION_ROOT,
   ROOT,
@@ -173,6 +206,7 @@ module.exports = {
   createWorker,
   updateStatus,
   appendHandoff,
+  completeHandoff,
   isLightweight,
   slugify,
   timestamp,

package/scripts/orchestration/session.cjs

@@ -11,11 +11,12 @@ const {
   createWorker,
   updateStatus,
   appendHandoff,
+  completeHandoff,
   slugify,
   relPosix
 } = require('./lib/orchestrator.cjs');
 
-const COMMANDS = ['init', 'add-worker', 'status', 'set-status', 'append', 'list', 'help'];
+const COMMANDS = ['init', 'add-worker', 'status', 'set-status', 'append', 'complete', 'list', 'help'];
 
 function usage() {
   console.log([
@@ -25,6 +26,9 @@ function usage() {
     '                                                       [--context <line>]... [--deliverable <line>]... [--lightweight]',
     '  node scripts/orchestration/session.js set-status <session> <worker> <state> [--details <text>]',
     '  node scripts/orchestration/session.js append <session> <worker> <section> --content <text>',
+    '  node scripts/orchestration/session.cjs complete <session> <worker>',
+    '                                                       [--summary <md>] [--files <md>]',
+    '                                                       [--validation <md>] [--follow-ups <md>]',
     '  node scripts/orchestration/session.js status <session>',
     '  node scripts/orchestration/session.js list',
     '',
@@ -110,6 +114,22 @@ function cmdAppend(session, worker, section, flags) {
   console.log(relPosix(file));
 }
 
+function cmdComplete(session, worker, flags) {
+  if (!session || !worker) throw new Error('session and worker names are required');
+  const fields = {
+    summary: flags.summary,
+    files: flags.files,
+    validation: flags.validation,
+    followUps: flags['follow-ups']
+  };
+  const provided = Object.values(fields).some(v => v !== undefined && v !== null);
+  if (!provided) {
+    throw new Error('at least one of --summary | --files | --validation | --follow-ups is required');
+  }
+  const file = completeHandoff(session, worker, fields);
+  console.log(relPosix(file));
+}
+
 function cmdStatus(session) {
   if (!session) throw new Error('session-name is required');
   const dir = sessionDir(session);
@@ -161,6 +181,9 @@ function main() {
       case 'append':
         cmdAppend(positional[0], positional[1], positional[2], flags);
         break;
+      case 'complete':
+        cmdComplete(positional[0], positional[1], flags);
+        break;
       case 'status':
         cmdStatus(positional[0]);
         break;

package/skills/ai-ml/fine-tuning-expert/SKILL.md

@@ -0,0 +1,162 @@
+---
+name: fine-tuning-expert
+description: "Use when fine-tuning LLMs, training custom models, or adapting foundation models for specific tasks. Invoke for configuring LoRA/QLoRA adapters, preparing JSONL training datasets, setting hyperparameters for fine-tuning runs, adapter training, transfer learning, finetuning with Hugging Face PEFT, OpenAI fine-tuning, instruction tuning, RLHF, DPO, or quantizing and deploying fine-tuned models. Trigger terms include: LoRA, QLoRA, PEFT, finetuning, fine-tuning, adapter tuning, LLM training, model training, custom model."
+license: MIT
+metadata:
+  author: https://github.com/Jeffallan
+  version: "1.1.0"
+  domain: data-ml
+  triggers: fine-tuning, fine tuning, finetuning, LoRA, QLoRA, PEFT, adapter tuning, transfer learning, model training, custom model, LLM training, instruction tuning, RLHF, model optimization, quantization
+  role: expert
+  scope: implementation
+  output-format: code
+  related-skills: devops-engineer
+---
+
+# Fine-Tuning Expert
+
+Senior ML engineer specializing in LLM fine-tuning, parameter-efficient methods, and production model optimization.
+
+## Core Workflow
+
+1. **Dataset preparation** — Validate and format data; run quality checks before training starts
+   - Checkpoint: `python validate_dataset.py --input data.jsonl` — fix all errors before proceeding
+2. **Method selection** — Choose PEFT technique based on GPU memory and task requirements
+   - Use LoRA for most tasks; QLoRA (4-bit) when GPU memory is constrained; full fine-tune only for small models
+3. **Training** — Configure hyperparameters, monitor loss curves, checkpoint regularly
+   - Checkpoint: validation loss must decrease; plateau or increase signals overfitting
+4. **Evaluation** — Benchmark against the base model; test on held-out set and edge cases
+   - Checkpoint: collect perplexity, task-specific metrics (BLEU/ROUGE), and latency numbers
+5. **Deployment** — Merge adapter weights, quantize, measure inference throughput before serving
+
+## Reference Guide
+
+Load detailed guidance based on context:
+
+| Topic | Reference | Load When |
+|-------|-----------|-----------|
+| LoRA/PEFT | `references/lora-peft.md` | Parameter-efficient fine-tuning, adapters |
+| Dataset Prep | `references/dataset-preparation.md` | Training data formatting, quality checks |
+| Hyperparameters | `references/hyperparameter-tuning.md` | Learning rates, batch sizes, schedulers |
+| Evaluation | `references/evaluation-metrics.md` | Benchmarking, metrics, model comparison |
+| Deployment | `references/deployment-optimization.md` | Model merging, quantization, serving |
+
+## Minimal Working Example — LoRA Fine-Tuning with Hugging Face PEFT
+
+```python
+from datasets import load_dataset
+from transformers import AutoTokenizer, AutoModelForCausalLM, TrainingArguments
+from peft import LoraConfig, get_peft_model, TaskType
+from trl import SFTTrainer
+import torch
+
+# 1. Load base model and tokenizer
+model_id = "meta-llama/Llama-3-8B"
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+tokenizer.pad_token = tokenizer.eos_token
+
+model = AutoModelForCausalLM.from_pretrained(
+    model_id,
+    torch_dtype=torch.bfloat16,
+    device_map="auto",
+)
+
+# 2. Configure LoRA adapter
+lora_config = LoraConfig(
+    task_type=TaskType.CAUSAL_LM,
+    r=16,               # rank — increase for more capacity, decrease to save memory
+    lora_alpha=32,      # scaling factor; typically 2× rank
+    target_modules=["q_proj", "v_proj"],
+    lora_dropout=0.05,
+    bias="none",
+)
+model = get_peft_model(model, lora_config)
+model.print_trainable_parameters()  # verify: should be ~0.1–1% of total params
+
+# 3. Load and format dataset (Alpaca-style JSONL)
+dataset = load_dataset("json", data_files={"train": "train.jsonl", "test": "test.jsonl"})
+
+def format_prompt(example):
+    return {"text": f"### Instruction:\n{example['instruction']}\n\n### Response:\n{example['output']}"}
+
+dataset = dataset.map(format_prompt)
+
+# 4. Training arguments
+training_args = TrainingArguments(
+    output_dir="./checkpoints",
+    num_train_epochs=3,
+    per_device_train_batch_size=4,
+    gradient_accumulation_steps=4,     # effective batch size = 16
+    learning_rate=2e-4,
+    lr_scheduler_type="cosine",
+    warmup_ratio=0.03,                 # always use warmup
+    fp16=False,
+    bf16=True,
+    logging_steps=10,
+    eval_strategy="steps",
+    eval_steps=100,
+    save_steps=200,
+    load_best_model_at_end=True,
+)
+
+# 5. Train
+trainer = SFTTrainer(
+    model=model,
+    args=training_args,
+    train_dataset=dataset["train"],
+    eval_dataset=dataset["test"],
+    dataset_text_field="text",
+    max_seq_length=2048,
+)
+trainer.train()
+
+# 6. Save adapter weights only
+model.save_pretrained("./lora-adapter")
+tokenizer.save_pretrained("./lora-adapter")
+```
+
+**QLoRA variant** — add these lines before loading the model to enable 4-bit quantization:
+```python
+from transformers import BitsAndBytesConfig
+
+bnb_config = BitsAndBytesConfig(
+    load_in_4bit=True,
+    bnb_4bit_quant_type="nf4",
+    bnb_4bit_compute_dtype=torch.bfloat16,
+    bnb_4bit_use_double_quant=True,
+)
+model = AutoModelForCausalLM.from_pretrained(model_id, quantization_config=bnb_config, device_map="auto")
+```
+
+**Merge adapter into base model for deployment:**
+```python
+from peft import PeftModel
+
+base = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.bfloat16)
+merged = PeftModel.from_pretrained(base, "./lora-adapter").merge_and_unload()
+merged.save_pretrained("./merged-model")
+```
+
+## Constraints
+
+### MUST DO
+- Validate dataset quality before training
+- Use parameter-efficient methods for large models (>7B)
+- Monitor training/validation loss curves
+- Document hyperparameters and training config
+- Version datasets and model checkpoints
+- Always include a learning rate warmup
+
+### MUST NOT DO
+- Skip data quality validation
+- Overfit on small datasets — use regularisation (dropout, weight decay) and early stopping
+- Merge incompatible adapters (mismatched rank, base model, or target modules)
+- Deploy without evaluation against a held-out set and latency benchmark
+
+## Output Templates
+
+When implementing fine-tuning, always provide:
+1. **Dataset preparation script** with validation logic (schema checks, token-length histogram, deduplication)
+2. **Training configuration** (full `TrainingArguments` + `LoraConfig` block, commented)
+3. **Evaluation script** reporting perplexity, task-specific metrics, and latency
+4. **Brief design rationale** — why this PEFT method, rank, and learning rate were chosen for this task