100+ AI Agents, Agent Skills and RAG Apps - Free and Open Source.
feat: update the self-improving agent skills
Shubhamsaboobd174d55d9f9
diff --git a/.gitignore b/.gitignoreindex 774ab181..47702b3c 100644--- a/.gitignore+++ b/.gitignore@@ -2,3 +2,15 @@__pycache__/*.pyc*.pyo++# Environment files+.env+.env.local++# Virtual environments+venv/+.venv/++# Node+node_modules/+.next/diff --git a/awesome_agent_skills/self-improving-agent-skills/README.md b/awesome_agent_skills/self-improving-agent-skills/README.mdindex b3c36c5b..cd030916 100644--- a/awesome_agent_skills/self-improving-agent-skills/README.md+++ b/awesome_agent_skills/self-improving-agent-skills/README.md@@ -1,31 +1,31 @@# Self-Improving Agent Skills-Automatically optimize your agent skills using a multi-agent system built with **Google ADK (Agent Development Kit)** and **Gemini 3**. Upload a skill, let the agents generate test scenarios and evaluation criteria, then watch as three specialized ADK agents collaborate to improve your skill through iterative optimization.+Automatically optimize your agent skills using a multi-agent system built with **Google ADK (Agent Development Kit)** and **Gemini**. Upload a skill, let the agents generate test scenarios and evaluation criteria, then watch as three specialized ADK agents collaborate to improve your skill through iterative optimization.## How It WorksThis app implements an automated skill improvement loop inspired by Karpathy's autoresearch methodology, powered by a team of ADK agents:1. **Upload**: Drop in your skill folder (following [agentskills.io](https://agentskills.io) spec)-2. **Configure**: AI generates test scenarios and evaluation criteria. Edit, add, or regenerate as needed-3. **Optimize**: Three ADK agents collaborate — one executes, one diagnoses failures, one applies fixes+2. **Configure**: The Executor agent generates test scenarios and evaluation criteria. Edit, add, or regenerate as needed+3. **Optimize**: Three ADK agents collaborate — one executes and scores, one diagnoses failures, one applies fixes4. **Results**: Download your improved skill with a detailed changelog### The ADK Agent Team| Agent | Role | What It Does ||-------|------|-------------|-| **Executor** | Skill Runner | Faithfully executes the skill against test scenarios, producing outputs exactly as the skill instructs |-| **Analyst** | Failure Diagnostician | Examines failed evaluations, identifies root causes, and recommends a specific mutation strategy |-| **Mutator** | Prompt Editor | Makes exactly ONE targeted change to the skill prompt based on the analyst's diagnosis |+| **Executor** | Skill Runner & Scorer | Executes the skill against test scenarios, scores outputs against evaluation criteria, and generates initial test scenarios during analysis |+| **Analyst** | Failure Diagnostician | Examines failed evaluations, identifies root causes, and recommends a mutation strategy. Uses Pydantic `output_schema` for guaranteed structured JSON |+| **Mutator** | Prompt Editor | Makes exactly ONE targeted change to the skill prompt based on the analyst's diagnosis. Uses Pydantic `output_schema` for guaranteed structured JSON |### The Optimization Loop- The **Executor** agent runs the skill against all test scenarios-- Outputs are scored using binary yes/no evaluation criteria+- The **Executor** then scores each output against binary yes/no evaluation criteria- The **Analyst** agent diagnoses failure patterns and picks a strategy (`add_example`, `add_constraint`, `restructure`, or `add_edge_case`)- The **Mutator** agent applies ONE surgical fix to the skill prompt-- The modified skill is re-tested+- The **Executor** re-runs and re-scores the modified skill- Changes are kept if the score improves, reverted if not- Repeats until the target pass rate is reached or max rounds hit@@ -36,8 +36,7 @@ self-improving-agent-skills/├── backend/ # FastAPI server + ADK optimization engine│ ├── app.py # REST API endpoints + SSE streaming│ ├── adk_optimizer.py # Multi-agent optimizer (Executor, Analyst, Mutator)-│ ├── requirements.txt-│ └── optimizer.py # Legacy single-model optimizer (unused)+│ └── requirements.txt├── frontend/ # Next.js + React + Tailwind│ ├── src/│ │ ├── app/ # Main page + layout@@ -52,9 +51,9 @@ self-improving-agent-skills/## Tech Stack-- **Backend**: Python 3.10+, FastAPI, Google ADK, google-genai SDK+- **Backend**: Python 3.10+, FastAPI, Google ADK, Pydantic- **Frontend**: Next.js 15, React 19, Tailwind CSS v4, Recharts-- **AI**: Google ADK multi-agent system with Gemini 3 Flash for execution, analysis, and mutation+- **AI**: Google ADK multi-agent system with Gemini (`gemini-3-flash-preview`) — structured output via `output_schema` on Analyst and Mutator agents- **Real-time**: Server-Sent Events (SSE) for live optimization progress## Quick Start@@ -71,9 +70,9 @@ source venv/bin/activate # On Windows: venv\Scripts\activate# Install dependenciespip install -r requirements.txt-# Set up environment (optional, can also pass via header)+# Set up environment (optional — the app will prompt for your API key in the UI)cp .env.example .env-# Edit .env and add your GEMINI_API_KEY+# Edit .env and add your GOOGLE_API_KEY# Run serverpython app.py@@ -151,22 +150,23 @@ Then upload the zip in the app.## How the Multi-Agent Optimization Works### 1. Analysis Phase-Gemini analyzes your skill and generates:+The **Executor** agent analyzes your skill and generates:- 3-4 diverse test scenarios- 4-6 binary evaluation criteria (yes/no questions)+You can edit, add, or remove scenarios and criteria before optimization begins.+### 2. Baseline Run-The **Executor** agent runs the skill against all scenarios. Each output is scored against all evaluation criteria. This establishes the starting score.+The **Executor** agent runs the skill against all scenarios and scores each output against all evaluation criteria. This establishes the starting score.### 3. Optimization LoopFor each round, the three agents collaborate:-1. **Executor** runs the skill against all test scenarios-2. Outputs are scored against evaluation criteria-3. **Analyst** examines failures, identifies root cause, and selects a mutation strategy-4. **Mutator** applies ONE specific change to improve the skill-5. **Executor** re-runs the modified skill-6. Score is compared — keep the change if improved, revert if not-7. Repeat until target pass rate or max rounds reached+1. **Executor** runs the skill against all test scenarios and scores the outputs+2. **Analyst** examines failures, identifies root cause, and selects a mutation strategy (returns structured JSON via `output_schema`)+3. **Mutator** applies ONE specific change to improve the skill (returns structured JSON via `output_schema`)+4. **Executor** re-runs and re-scores the modified skill+5. Score is compared — keep the change if improved, revert if not+6. Repeat until target pass rate or max rounds reached### 4. Output- Improved SKILL.md with all successful changes applied@@ -175,36 +175,47 @@ For each round, the three agents collaborate:## API Endpoints-- `POST /api/upload` - Upload skill zip file-- `POST /api/upload-files` - Upload multiple files (folder upload)-- `POST /api/analyze` - Generate scenarios and evals (requires Gemini API key)-- `POST /api/regenerate` - Regenerate scenarios and evals-- `POST /api/update-config` - Save user's selected/edited config-- `POST /api/start/{session_id}` - Start optimization-- `GET /api/stream/{session_id}` - SSE stream of optimization progress-- `POST /api/stop/{session_id}` - Stop optimization-- `GET /api/download/{session_id}` - Download improved skill-- `GET /api/examples` - List available example skills-- `POST /api/examples/{name}/load` - Load an example skill-- `GET /api/status/{session_id}` - Poll-based status endpoint+| Method | Endpoint | Description |+|--------|----------|-------------|+| `POST` | `/api/upload` | Upload skill zip file (max 10MB, text files only) |+| `POST` | `/api/upload-files` | Upload multiple files (folder upload) |+| `POST` | `/api/analyze` | Generate scenarios and evals (requires Gemini API key) |+| `POST` | `/api/regenerate` | Regenerate scenarios and evals |+| `POST` | `/api/update-config` | Save user's selected/edited config |+| `POST` | `/api/start/{session_id}` | Start optimization |+| `GET` | `/api/stream/{session_id}` | SSE stream of optimization progress |+| `POST` | `/api/stop/{session_id}` | Stop optimization |+| `GET` | `/api/download/{session_id}` | Download improved skill |+| `GET` | `/api/examples` | List available example skills |+| `POST` | `/api/examples/{name}/load` | Load an example skill |+| `GET` | `/api/status/{session_id}` | Poll-based status endpoint |+| `GET` | `/health` | Health check |## Configuration### Backend-Set `GEMINI_API_KEY` in `.env` or pass via request header. Server runs on port **8891**.+The Gemini API key is passed from the frontend with each request. Optionally set `GOOGLE_API_KEY` in `.env` for local development. Server runs on port **8891**.++Upload limits:+- **10MB** max total upload size+- **1MB** max per file+- **50** max files per upload+- Text files only (`.md`, `.txt`, `.json`, `.yaml`, `.py`, `.js`, `.ts`, etc.)++Sessions expire after **1 hour** automatically.### Frontend-API key is stored in component state (not persisted) and sent with each request.+API key is entered in the UI, stored in component state (not persisted), and sent with each request. The key is passed to the backend which sets `GOOGLE_API_KEY` for ADK agent authentication.### Optimization Parameters-In `RunningStep.tsx`, adjust `max_rounds`:+In `RunningStep.tsx`, adjust `max_rounds` (capped at 50):```typescriptbody: JSON.stringify({- max_rounds: 20, // Default: 20+ max_rounds: 20, // Default: 20, max: 50}),```@@ -238,4 +249,4 @@ Both servers support hot reload. Edit code and see changes immediately.This tool applies Andrej Karpathy's autoresearch methodology (using LLMs to iteratively improve their own prompts) to agent skills. The key insight: rather than manually tweaking prompts, define success criteria and let the AI optimize itself — now powered by a team of specialized ADK agents.-Original concept: [https://github.com/karpathy/autoresearch](https://github.com/karpathy/autoresearch)\ No newline at end of file+Original concept: [https://github.com/karpathy/autoresearch](https://github.com/karpathy/autoresearch)diff --git a/awesome_agent_skills/self-improving-agent-skills/backend/adk_optimizer.py b/awesome_agent_skills/self-improving-agent-skills/backend/adk_optimizer.pyindex 1b815649..ce12d6fc 100644--- a/awesome_agent_skills/self-improving-agent-skills/backend/adk_optimizer.py+++ b/awesome_agent_skills/self-improving-agent-skills/backend/adk_optimizer.py@@ -1,26 +1,48 @@"""Multi-Agent Skill Optimizer using Google ADK.3 ADK agents work together to improve agent skills:- Executor: runs the skill against test scenarios+ Executor: runs the skill against test scenarios, scores outputs, analyzes skillsAnalyst: diagnoses why evals failed, picks a mutation strategyMutator: makes one targeted fix per round"""import json-import re-import uuid-from typing import Callable, Optional+import os+from typing import Callable, List, Optional-from google.adk import Agent+from pydantic import BaseModel, Field++from google.adk.agents import Agentfrom google.adk.runners import Runnerfrom google.adk.sessions import InMemorySessionService-from google import genaifrom google.genai import types+# -- Pydantic schemas for structured agent output ----------------------------+++class FailureAnalysis(BaseModel):+ diagnosis: str = Field(description="Root cause of failures")+ mutation_strategy: str = Field(+ description="One of: add_example, add_constraint, restructure, add_edge_case"+ )+ target_section: str = Field(description="Which part of the skill to change")+ suggested_change: str = Field(description="What specific change to make")+++class SkillMutation(BaseModel):+ description: str = Field(description="Short description of the change made")+ reasoning: str = Field(description="Why this change should help")+ new_skill_md: str = Field(description="The full updated SKILL.md content")+++# -- Optimizer ---------------------------------------------------------------++class SkillOptimizer:def __init__(self, api_key: str, model: str = "gemini-3-flash-preview"):- self.client = genai.Client(api_key=api_key)+ # ADK agents authenticate via this env var+ os.environ["GOOGLE_API_KEY"] = api_keyself.model = modelself._session_service = InMemorySessionService()self._call_id = 0@@ -29,9 +51,14 @@ class SkillOptimizer:name="executor",model=model,instruction=(- "You execute agent skills faithfully. Given a skill's instructions "- "and a user request, produce the output that skill would generate. "- "Follow the instructions exactly. No meta-commentary."+ "You are a versatile skill execution agent. You have three modes:\n\n"+ "1. EXECUTE MODE: Given a skill's instructions and a user request, "+ "produce the output that skill would generate. Follow the instructions "+ "exactly. No meta-commentary.\n\n"+ "2. ANALYZE MODE: Given a skill definition, generate test scenarios "+ "and evaluation criteria. Return valid JSON.\n\n"+ "3. SCORE MODE: Given an output and evaluation criteria, score the "+ "output against each criterion. Return valid JSON."),)self.analyst = Agent(@@ -40,22 +67,23 @@ class SkillOptimizer:instruction=("You diagnose why agent skill evaluations fail. ""Given failed eval results, identify the root cause and suggest "- "a specific fix. Return valid JSON with keys: "- "diagnosis, mutation_strategy (add_example|add_constraint|restructure|add_edge_case), "- "target_section, suggested_change."+ "a specific fix. Pick one mutation_strategy from: "+ "add_example, add_constraint, restructure, or add_edge_case."),+ output_schema=FailureAnalysis,)self.mutator = Agent(name="mutator",model=model,instruction=("You edit agent skill files. Given a SKILL.md and a diagnosis, "- "make exactly ONE targeted change. Keep structure and frontmatter intact. "- "Return valid JSON with keys: description, reasoning, new_skill_md."+ "make exactly ONE targeted change. Keep the YAML frontmatter and "+ "overall structure intact. Return the complete updated SKILL.md."),+ output_schema=SkillMutation,)- # -- Agent runner helper --------------------------------------------------+ # -- Agent runner helpers ------------------------------------------------async def _ask(self, agent: Agent, prompt: str) -> str:"""Run an ADK agent with a prompt, return text response."""@@ -79,32 +107,26 @@ class SkillOptimizer:text += part.textreturn text- # -- Gemini direct call (for structured JSON) ------------------------------- async def _gen_json(self, prompt: str):- """Call Gemini directly for structured JSON output."""- resp = await self.client.aio.models.generate_content(- model=self.model,- contents=prompt,- config=types.GenerateContentConfig(- response_mime_type="application/json"- ),- )- return json.loads(resp.text)-- # -- Parse JSON from agent text (agents don't guarantee pure JSON) ---------- def _parse_json(self, text: str, fallback: dict) -> dict:+ async def _ask_json(self, agent: Agent, prompt: str, fallback=None):+ """Run an ADK agent and parse the JSON response."""+ text = await self._ask(agent, prompt)try:return json.loads(text)except json.JSONDecodeError:- m = re.search(r"\{.*\}", text, re.DOTALL)- if m:+ # Try raw_decode to handle extra data after valid JSON+ decoder = json.JSONDecoder()+ idx = text.find("{")+ if idx == -1:+ idx = text.find("[")+ if idx != -1:try:- return json.loads(m.group(0))+ result, _ = decoder.raw_decode(text, idx)+ return resultexcept json.JSONDecodeError:pass- return fallback+ if fallback is not None:+ return fallback+ raise# -- Public API -----------------------------------------------------------@@ -120,26 +142,21 @@ class SkillOptimizer:f"## {k}\n{v}" for k, v in refs.items())- prompt = f"""Analyze this agent skill and generate test scenarios with evaluation criteria.--# SKILL.md-{skill_md}-{ref_text}--Generate:-1. 3-4 diverse test scenarios (realistic user inputs that test different aspects)-2. 4-6 binary yes/no evaluation criteria--Return JSON:-{{- "scenarios": [- {{"id": 1, "name": "short name", "description": "short name", "input": "the user request to test"}}- ],- "evals": [- {{"id": 1, "name": "what to check", "criterion": "what to check", "question": "yes/no question about the output", "pass_condition": "what yes looks like", "fail_condition": "what no looks like"}}- ]-}}"""- return await self._gen_json(prompt)+ prompt = (+ f"Analyze this agent skill and generate test scenarios "+ f"with evaluation criteria.\n\n"+ f"# SKILL.md\n{skill_md}\n{ref_text}\n\n"+ f"Generate:\n"+ f"1. 3-4 diverse test scenarios (realistic user inputs)\n"+ f"2. 4-6 binary yes/no evaluation criteria\n\n"+ f"Return JSON:\n"+ f'{{"scenarios": [{{"id": 1, "name": "short name", "description": "short name", '+ f'"input": "the user request to test"}}], '+ f'"evals": [{{"id": 1, "name": "what to check", "criterion": "what to check", '+ f'"question": "yes/no question about the output", '+ f'"pass_condition": "what yes looks like", "fail_condition": "what no looks like"}}]}}'+ )+ return await self._ask_json(self.executor, prompt)async def optimize(self,@@ -251,32 +268,31 @@ Return JSON:# -- Internal helpers -----------------------------------------------------async def _score_skill(self, skill_md, scenarios, evals):- """Run Executor on all scenarios, then score with Gemini."""+ """Executor runs all scenarios, then scores outputs."""all_results = []total_passed = 0total_checks = 0per_eval = {e["id"]: {"passed": 0, "total": 0} for e in evals}for sc in scenarios:- # Executor runs the skill+ # Executor runs the skill (free-form text)output = await self._ask(self.executor,f"Execute this skill:\n\n{skill_md}\n\nUser request:\n{sc['input']}",)- # Score with Gemini (structured JSON, not an agent)- scores = await self._gen_json(- f"""Evaluate this output against criteria. Return JSON array.--Input: {sc['input']}-Output: {output}--Criteria:-{json.dumps(evals, indent=2)}--Return: [{{"eval_id": 1, "passed": true, "reason": "..."}}]"""+ # Executor scores the output (JSON)+ scoring = await self._ask_json(+ self.executor,+ (+ f"Evaluate this output against the criteria.\n\n"+ f"Input: {sc['input']}\n\n"+ f"Output: {output}\n\n"+ f"Criteria:\n{json.dumps(evals, indent=2)}\n\n"+ f"Return JSON: {{\"results\": [{{\"eval_id\": 1, \"passed\": true, \"reason\": \"...\"}}]}}"+ ),+ fallback={"results": []},)- if not isinstance(scores, list):- scores = scores.get("results", scores.get("evaluations", []))+ scores = scoring.get("results", []) if isinstance(scoring, dict) else scoringfor s in scores:eid = s.get("eval_id")@@ -304,40 +320,46 @@ Return: [{{"eval_id": 1, "passed": true, "reason": "..."}}]""""""Analyst agent diagnoses the worst failures."""failed = [d for d in details if not d.get("passed")]if not failed:- return {"diagnosis": "All passed", "mutation_strategy": "add_constraint",- "target_section": "N/A", "suggested_change": "none"}+ return {+ "diagnosis": "All passed",+ "mutation_strategy": "add_constraint",+ "target_section": "N/A",+ "suggested_change": "none",+ }- prompt = (- f"Diagnose these failures and suggest ONE fix.\n\n"- f"Skill:\n{skill_md[:2000]}\n\n"- f"Failed evals:\n{json.dumps(failed[:5], indent=2)}\n\n"- f"Return JSON: {{diagnosis, mutation_strategy, target_section, suggested_change}}"+ return await self._ask_json(+ self.analyst,+ (+ f"Diagnose these failures and suggest ONE fix.\n\n"+ f"Skill:\n{skill_md[:2000]}\n\n"+ f"Failed evals:\n{json.dumps(failed[:5], indent=2)}"+ ),+ fallback={+ "diagnosis": "Unable to diagnose",+ "mutation_strategy": "add_constraint",+ "target_section": "unknown",+ "suggested_change": "unclear",+ },)- text = await self._ask(self.analyst, prompt)- return self._parse_json(text, {- "diagnosis": text[:200],- "mutation_strategy": "add_constraint",- "target_section": "unknown",- "suggested_change": "unclear",- })async def _mutate_skill(self, skill_md, analysis):"""Mutator agent makes one targeted change."""- prompt = (- f"Apply this fix to the skill. Make ONE change only.\n\n"- f"SKILL.md:\n{skill_md}\n\n"- f"Diagnosis: {analysis.get('diagnosis')}\n"- f"Strategy: {analysis.get('mutation_strategy')}\n"- f"Target: {analysis.get('target_section')}\n"- f"Change: {analysis.get('suggested_change')}\n\n"- f"Return JSON: {{description, reasoning, new_skill_md}}"+ return await self._ask_json(+ self.mutator,+ (+ f"Apply this fix to the skill. Make ONE change only.\n\n"+ f"SKILL.md:\n{skill_md}\n\n"+ f"Diagnosis: {analysis.get('diagnosis')}\n"+ f"Strategy: {analysis.get('mutation_strategy')}\n"+ f"Target: {analysis.get('target_section')}\n"+ f"Change: {analysis.get('suggested_change')}"+ ),+ fallback={+ "description": "Failed to parse mutation",+ "reasoning": "",+ "new_skill_md": skill_md,+ },)- text = await self._ask(self.mutator, prompt)- return self._parse_json(text, {- "description": "Failed to parse mutation",- "reasoning": "",- "new_skill_md": skill_md,- })@staticmethoddef _strategy_stats(mutation_log):diff --git a/awesome_agent_skills/self-improving-agent-skills/backend/app.py b/awesome_agent_skills/self-improving-agent-skills/backend/app.pyindex 7b17e83c..2a4c058b 100644--- a/awesome_agent_skills/self-improving-agent-skills/backend/app.py+++ b/awesome_agent_skills/self-improving-agent-skills/backend/app.py@@ -1,8 +1,9 @@from fastapi import FastAPI, UploadFile, File, HTTPException, Formfrom fastapi.responses import StreamingResponse, FileResponsefrom fastapi.middleware.cors import CORSMiddleware-from pydantic import BaseModel+from pydantic import BaseModel, Fieldfrom typing import Optional, List, Dict+import timeimport uuidimport zipfileimport io@@ -19,6 +20,15 @@ from adk_optimizer import SkillOptimizerlogging.basicConfig(level=logging.INFO)logger = logging.getLogger(__name__)+MAX_UPLOAD_SIZE = 10 * 1024 * 1024 # 10 MB total+MAX_FILE_SIZE = 1 * 1024 * 1024 # 1 MB per file+MAX_FILE_COUNT = 50+SESSION_TTL = 3600 # 1 hour+ALLOWED_EXTENSIONS = {+ ".md", ".txt", ".json", ".yaml", ".yml", ".py", ".js", ".ts",+ ".html", ".css", ".xml", ".toml", ".cfg", ".ini", ".sh",+}+app = FastAPI()app.add_middleware(@@ -50,7 +60,7 @@ class RegenerateRequest(BaseModel):class StartRequest(BaseModel):gemini_api_key: str- max_rounds: Optional[int] = 20+ max_rounds: Optional[int] = Field(default=20, gt=0, le=50)def parse_skill_frontmatter(content: str) -> dict:@@ -95,16 +105,30 @@ def create_session_from_files(skill_files: dict, file_list: list) -> dict:"changelog": [],"current_skill_md": skill_md,"original_skill_md": skill_md,+ "created_at": time.time(),}return {"session_id": session_id, "file_list": file_list, "metadata": metadata}+def _is_allowed_file(name: str) -> bool:+ """Check if file extension is in the allowed text-file list."""+ _, ext = os.path.splitext(name)+ return ext.lower() in ALLOWED_EXTENSIONS+++def _is_safe_path(name: str) -> bool:+ """Reject path traversal attempts."""+ return ".." not in name and not os.path.isabs(name)++@app.post("/api/upload")async def upload_skill(file: UploadFile = File(...)):"""Accept zip file or multiple files, extract, return file list + parsed SKILL.md metadata"""if not file.filename.endswith(".zip"):raise HTTPException(status_code=400, detail="Only .zip files are accepted")content = await file.read()+ if len(content) > MAX_UPLOAD_SIZE:+ raise HTTPException(status_code=413, detail=f"Upload exceeds {MAX_UPLOAD_SIZE // (1024*1024)}MB limit")try:with zipfile.ZipFile(io.BytesIO(content)) as zf:skill_files = {}@@ -112,8 +136,20 @@ async def upload_skill(file: UploadFile = File(...)):for name in zf.namelist():if name.endswith("/") or name.startswith("__MACOSX") or "/.DS_Store" in name or name.endswith(".DS_Store"):continue- file_content = zf.read(name).decode("utf-8", errors="ignore")- # Strip leading directory if all files share a common prefix+ if not _is_safe_path(name):+ logger.warning(f"Skipping unsafe zip entry: {name}")+ continue+ if not _is_allowed_file(name):+ logger.info(f"Skipping non-text file: {name}")+ continue+ raw = zf.read(name)+ if len(raw) > MAX_FILE_SIZE:+ logger.warning(f"Skipping oversized file: {name} ({len(raw)} bytes)")+ continue+ if len(file_list) >= MAX_FILE_COUNT:+ logger.warning("Max file count reached, skipping remaining entries")+ break+ file_content = raw.decode("utf-8", errors="ignore")skill_files[name] = file_contentfile_list.append(name)@@ -130,19 +166,35 @@ async def upload_skill(file: UploadFile = File(...)):except HTTPException:raiseexcept Exception as e:- raise HTTPException(status_code=500, detail=f"Error processing file: {str(e)}")+ logger.error(f"Upload error: {traceback.format_exc()}")+ raise HTTPException(status_code=500, detail="Error processing file")@app.post("/api/upload-files")async def upload_files(files: List[UploadFile] = File(...)):"""Accept multiple files (folder upload via webkitdirectory)"""+ if len(files) > MAX_FILE_COUNT:+ raise HTTPException(status_code=413, detail=f"Too many files (max {MAX_FILE_COUNT})")skill_files = {}file_list = []+ total_size = 0for f in files:if f.filename.startswith(".") or "/.DS_Store" in (f.filename or "") or "__MACOSX" in (f.filename or ""):continue- content = await f.read()name = f.filename or "unknown"+ if not _is_safe_path(name):+ logger.warning(f"Skipping unsafe path: {name}")+ continue+ if not _is_allowed_file(name):+ logger.info(f"Skipping non-text file: {name}")+ continue+ content = await f.read()+ total_size += len(content)+ if total_size > MAX_UPLOAD_SIZE:+ raise HTTPException(status_code=413, detail=f"Total upload exceeds {MAX_UPLOAD_SIZE // (1024*1024)}MB limit")+ if len(content) > MAX_FILE_SIZE:+ logger.warning(f"Skipping oversized file: {name} ({len(content)} bytes)")+ continueskill_files[name] = content.decode("utf-8", errors="ignore")file_list.append(name)@@ -170,7 +222,8 @@ async def analyze_skill(request: AnalyzeRequest):session["status"] = "analyzed"return {"scenarios": analysis["scenarios"], "evals": analysis["evals"]}except Exception as e:- raise HTTPException(status_code=500, detail=f"Analysis failed: {str(e)}")+ logger.error(f"Analysis error: {traceback.format_exc()}")+ raise HTTPException(status_code=500, detail="Analysis failed. Check your API key and try again.")@app.post("/api/regenerate")@@ -409,8 +462,12 @@ async def list_examples():@app.post("/api/examples/{example_name}/load")async def load_example(example_name: str):"""Load an example skill as if it were uploaded"""- examples_dir = os.path.join(os.path.dirname(__file__), "..", "example_skills")- skill_dir = os.path.join(examples_dir, example_name)+ if not re.fullmatch(r"[a-zA-Z0-9_-]+", example_name):+ raise HTTPException(status_code=400, detail="Invalid example name")+ examples_dir = os.path.realpath(os.path.join(os.path.dirname(__file__), "..", "example_skills"))+ skill_dir = os.path.realpath(os.path.join(examples_dir, example_name))+ if not skill_dir.startswith(examples_dir + os.sep):+ raise HTTPException(status_code=400, detail="Invalid example name")if not os.path.isdir(skill_dir):raise HTTPException(status_code=404, detail="Example skill not found")skill_files = {}@@ -446,6 +503,27 @@ async def health_check():return {"status": "healthy"}+async def _cleanup_expired_sessions():+ """Periodically remove sessions older than SESSION_TTL."""+ while True:+ await asyncio.sleep(300) # every 5 minutes+ now = time.time()+ expired = [+ sid for sid, s in sessions.items()+ if now - s.get("created_at", now) > SESSION_TTL+ and s.get("status") not in ("running",)+ ]+ for sid in expired:+ del sessions[sid]+ if expired:+ logger.info(f"Cleaned up {len(expired)} expired session(s)")++[email protected]_event("startup")+async def startup():+ asyncio.create_task(_cleanup_expired_sessions())++if __name__ == "__main__":import uvicornuvicorn.run(app, host="0.0.0.0", port=8891)diff --git a/awesome_agent_skills/self-improving-agent-skills/backend/optimizer.py b/awesome_agent_skills/self-improving-agent-skills/backend/optimizer.pydeleted file mode 100644index f5fbc204..00000000--- a/awesome_agent_skills/self-improving-agent-skills/backend/optimizer.py+++ /dev/null@@ -1,419 +0,0 @@-from google import genai-from google.genai import types-import json-import asyncio-import copy-import re-from typing import Dict, List, Optional, Callable---class SkillOptimizer:- def __init__(self, api_key: str, model: str = "gemini-3-flash-preview"):- self.client = genai.Client(api_key=api_key)- self.model = model-- async def analyze_skill(self, skill_files: Dict[str, str]) -> dict:- """Read all skill files, generate 3-5 test scenarios + 3-6 binary eval criteria.-- Returns:- {"scenarios": [{"id": 1, "input": "...", "description": "..."}],- "evals": [{"id": 1, "name": "...", "question": "...", "pass_condition": "...", "fail_condition": "..."}]}- """- skill_md = skill_files.get("SKILL.md", "")- references = {k: v for k, v in skill_files.items() if k != "SKILL.md"}-- context = f"# Skill Content\n\n{skill_md}\n\n"- if references:- context += "# Reference Files\n\n"- for filename, content in references.items():- context += f"## {filename}\n\n{content}\n\n"-- prompt = f"""Analyze this agent skill and generate test scenarios and evaluation criteria.--{context}--Generate:-1. 3-5 diverse test scenarios that exercise different aspects of the skill-2. 3-6 binary evaluation criteria (yes/no questions) to measure quality--For each scenario, provide:-- id (number)-- description (what this scenario tests)-- input (the actual test input to send to the skill)--For each evaluation criterion, provide:-- id (number)-- name (short label)-- question (a yes/no question to evaluate the output)-- pass_condition (what makes it pass)-- fail_condition (what makes it fail)--Return valid JSON with this structure:-{{- "scenarios": [- {{"id": 1, "description": "...", "input": "..."}}- ],- "evals": [- {{"id": 1, "name": "...", "question": "...", "pass_condition": "...", "fail_condition": "..."}}- ]-}}"""-- response = await self.client.aio.models.generate_content(- model=self.model,- contents=prompt,- config=types.GenerateContentConfig(- response_mime_type="application/json"- )- )-- result = json.loads(response.text)- return result-- async def execute_skill(self, skill_md: str, test_input: str) -> str:- """Run the skill as a system prompt with the test input as user message.-- Returns:- The output from executing the skill.- """- response = await self.client.aio.models.generate_content(- model=self.model,- contents=[- types.Content(- role="user",- parts=[types.Part(text=f"System instructions:\n\n{skill_md}\n\n---\n\nUser request:\n\n{test_input}")]- )- ]- )-- return response.text-- async def score_output(self, output: str, test_input: str, evals: List[dict]) -> List[dict]:- """Score one output against all evals.-- Returns:- [{"eval_id": 1, "passed": true/false, "reason": "..."}]- """- eval_questions = "\n".join([- f"{i+1}. {ev['question']}\n Pass: {ev['pass_condition']}\n Fail: {ev['fail_condition']}"- for i, ev in enumerate(evals)- ])-- prompt = f"""Evaluate this output against the criteria below.--# Test Input-{test_input}--# Output-{output}--# Evaluation Criteria-{eval_questions}--For each criterion, determine if it passes (yes) or fails (no), and explain why.--Return valid JSON array:-[- {{"eval_id": 1, "passed": true, "reason": "..."}},- {{"eval_id": 2, "passed": false, "reason": "..."}}-]"""-- response = await self.client.aio.models.generate_content(- model=self.model,- contents=prompt,- config=types.GenerateContentConfig(- response_mime_type="application/json"- )- )-- return json.loads(response.text)-- async def analyze_failures(self, experiment_results: list, evals: list, changelog: list) -> str:- """Analyze which evals fail most and why.-- Returns:- Text analysis of failure patterns.- """- results_summary = json.dumps(experiment_results, indent=2)- evals_summary = json.dumps(evals, indent=2)- changelog_summary = json.dumps(changelog, indent=2)-- prompt = f"""Analyze these experiment results to identify failure patterns.--# Evaluation Criteria-{evals_summary}--# Recent Experiment Results-{results_summary}--# Previous Changes (if any)-{changelog_summary}--Analyze:-1. Which evaluation criteria fail most frequently?-2. What patterns do you see in the failures?-3. What specific aspect of the skill needs improvement?--Provide a concise analysis focusing on the root cause."""-- response = await self.client.aio.models.generate_content(- model=self.model,- contents=prompt- )-- return response.text-- async def mutate_skill(self, current_skill_md: str, failure_analysis: str, changelog: list) -> dict:- """Propose ONE targeted change.-- Returns:- {"description": "what changed", "reasoning": "why", "new_skill_md": "..."}- """- changelog_summary = "\n".join([- f"- {entry['description']} ({entry['status']})"- for entry in changelog[-5:] if changelog- ]) if changelog else "No previous changes"-- prompt = f"""You are improving an agent skill prompt. Based on the failure analysis, propose ONE specific change.--# Current Skill Prompt-{current_skill_md}--# Failure Analysis-{failure_analysis}--# Recent Changes-{changelog_summary}--Propose ONE targeted change to address the failures. Changes can include:-- Adding specific rules or constraints-- Clarifying existing instructions-- Adding examples-- Restructuring sections-- Adding checklists or formats--Be surgical: change only what's needed. Avoid changes that were already tried and failed.--Return valid JSON:-{{- "description": "Brief description of the change",- "reasoning": "Why this addresses the failures",- "new_skill_md": "The complete modified SKILL.md content"-}}"""-- response = await self.client.aio.models.generate_content(- model=self.model,- contents=prompt,- config=types.GenerateContentConfig(- response_mime_type="application/json"- )- )-- return json.loads(response.text)-- async def run_experiment(self, skill_md: str, scenarios: list, evals: list, runs_per_scenario: int = 2) -> dict:- """Run skill across all scenarios, score all outputs.-- Returns:- {"score": 14, "max_score": 20, "pass_rate": 70.0, "per_eval": [...], "per_scenario": [...]}- """- all_results = []- per_scenario_results = []-- for scenario in scenarios:- scenario_results = []-- for run_idx in range(runs_per_scenario):- output = await self.execute_skill(skill_md, scenario["input"])- scores = await self.score_output(output, scenario["input"], evals)-- scenario_results.append({- "scenario_id": scenario["id"],- "run": run_idx + 1,- "output": output,- "scores": scores- })- all_results.extend(scores)-- passed = sum(1 for r in scenario_results for s in r["scores"] if s["passed"])- total = len(scenario_results) * len(evals)-- per_scenario_results.append({- "scenario_id": scenario["id"],- "description": scenario["description"],- "passed": passed,- "total": total,- "pass_rate": round((passed / total * 100), 1) if total > 0 else 0- })-- per_eval_results = []- for eval_item in evals:- eval_scores = [r for r in all_results if r["eval_id"] == eval_item["id"]]- passed = sum(1 for s in eval_scores if s["passed"])- total = len(eval_scores)-- per_eval_results.append({- "eval_id": eval_item["id"],- "name": eval_item["name"],- "passed": passed,- "total": total,- "pass_rate": round((passed / total * 100), 1) if total > 0 else 0- })-- total_passed = sum(1 for r in all_results if r["passed"])- total_checks = len(all_results)-- return {- "score": total_passed,- "max_score": total_checks,- "pass_rate": round((total_passed / total_checks * 100), 1) if total_checks > 0 else 0,- "per_eval": per_eval_results,- "per_scenario": per_scenario_results,- "raw_results": all_results- }-- async def optimize(- self,- skill_files: dict,- scenarios: list,- evals: list,- max_rounds: int = 20,- callback: Optional[Callable] = None- ):- """Main optimization loop.-- 1. Run baseline experiment (experiment 0)- 2. Loop:- a. Analyze failures- b. Mutate skill (one change)- c. Run experiment- d. If score improved: keep. Else: revert.- e. Log to changelog- f. Call callback with progress event- g. Stop if 95%+ for 3 consecutive or max_rounds hit- 3. Return final results- """- skill_md = skill_files.get("SKILL.md", "")- current_skill_md = skill_md- original_skill_md = skill_md-- changelog = []- experiments = []- consecutive_95_plus = 0-- if callback:- await callback({"type": "status", "data": {"message": "Running baseline experiment..."}})-- baseline = await self.run_experiment(current_skill_md, scenarios, evals)- baseline_result = {- "experiment_id": 0,- "description": "Baseline",- "status": "baseline",- "score": baseline["score"],- "max_score": baseline["max_score"],- "pass_rate": baseline["pass_rate"],- "per_eval": baseline["per_eval"],- "per_scenario": baseline["per_scenario"]- }- experiments.append(baseline_result)-- if callback:- await callback({"type": "baseline", "data": baseline_result})-- best_score = baseline["score"]- best_skill_md = current_skill_md-- if baseline["pass_rate"] >= 95:- consecutive_95_plus = 1-- for round_idx in range(max_rounds):- if consecutive_95_plus >= 3:- break-- recent_experiments = experiments[-3:] if len(experiments) >= 3 else experiments- failure_analysis = await self.analyze_failures(recent_experiments, evals, changelog)-- if callback:- await callback({- "type": "experiment_start",- "data": {- "experiment_id": round_idx + 1,- "analyzing": True- }- })-- mutation = await self.mutate_skill(current_skill_md, failure_analysis, changelog)-- if callback:- await callback({- "type": "experiment_start",- "data": {- "experiment_id": round_idx + 1,- "description": mutation["description"]- }- })-- new_skill_md = mutation["new_skill_md"]-- result = await self.run_experiment(new_skill_md, scenarios, evals)-- if result["score"] > best_score:- status = "keep"- current_skill_md = new_skill_md- best_score = result["score"]- best_skill_md = new_skill_md-- if result["pass_rate"] >= 95:- consecutive_95_plus += 1- else:- consecutive_95_plus = 0- else:- status = "discard"- if result["pass_rate"] >= 95:- consecutive_95_plus += 1- else:- consecutive_95_plus = 0-- experiment_result = {- "experiment_id": round_idx + 1,- "description": mutation["description"],- "reasoning": mutation["reasoning"],- "status": status,- "score": result["score"],- "max_score": result["max_score"],- "pass_rate": result["pass_rate"],- "per_eval": result["per_eval"],- "per_scenario": result["per_scenario"]- }-- experiments.append(experiment_result)- changelog.append({- "experiment_id": round_idx + 1,- "description": mutation["description"],- "reasoning": mutation["reasoning"],- "status": status,- "score_before": best_score if status == "discard" else experiments[-2]["score"],- "score_after": result["score"]- })-- if callback:- await callback({"type": "experiment_result", "data": experiment_result})-- kept_count = sum(1 for e in experiments if e.get("status") == "keep")- discarded_count = sum(1 for e in experiments if e.get("status") == "discard")-- final_result = {- "baseline_score": baseline["pass_rate"],- "final_score": experiments[-1]["pass_rate"],- "experiments_run": len(experiments) - 1,- "kept": kept_count,- "discarded": discarded_count,- "changelog": changelog,- "experiments": experiments,- "improved_skill_md": best_skill_md,- "original_skill_md": original_skill_md- }-- if callback:- await callback({"type": "complete", "data": final_result})-- return final_resultdiff --git a/awesome_agent_skills/self-improving-agent-skills/backend/requirements.txt b/awesome_agent_skills/self-improving-agent-skills/backend/requirements.txtindex eec3bbd3..c2213f43 100644--- a/awesome_agent_skills/self-improving-agent-skills/backend/requirements.txt+++ b/awesome_agent_skills/self-improving-agent-skills/backend/requirements.txt@@ -1,6 +1,6 @@-fastapi-uvicorn[standard]-google-genai-google-adk-python-multipart-sse-starlette+fastapi>=0.115.0+uvicorn[standard]>=0.32.0+google-genai>=1.0.0+google-adk>=1.0.0+pydantic>=2.0.0+python-multipart>=0.0.18diff --git a/awesome_agent_skills/self-improving-agent-skills/frontend/package-lock.json b/awesome_agent_skills/self-improving-agent-skills/frontend/package-lock.jsonindex df113c2b..6039e066 100644--- a/awesome_agent_skills/self-improving-agent-skills/frontend/package-lock.json+++ b/awesome_agent_skills/self-improving-agent-skills/frontend/package-lock.json@@ -574,15 +574,15 @@}},"node_modules/@next/env": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/env/-/env-15.5.13.tgz",- "integrity": "sha512-6h7Fm29+/u1WBPcPaQl0xBov7KXB6i0c8oFlSlehD+PuZJQjzXQBuYzfkM32G5iWOlKsXXyRtcMaaqwspRBujA==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/env/-/env-15.5.15.tgz",+ "integrity": "sha512-vcmyu5/MyFzN7CdqRHO3uHO44p/QPCZkuTUXroeUmhNP8bL5PHFEhik22JUazt+CDDoD6EpBYRCaS2pISL+/hg==","license": "MIT"},"node_modules/@next/swc-darwin-arm64": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/swc-darwin-arm64/-/swc-darwin-arm64-15.5.13.tgz",- "integrity": "sha512-XrBbj2iY1mQSsJ8RoFClNpUB9uuZejP94v9pJuSAzdzwFVHeP+Vu2vzBCHwSObozgYNuTVwKhLukG1rGCgj8xA==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/swc-darwin-arm64/-/swc-darwin-arm64-15.5.15.tgz",+ "integrity": "sha512-6PvFO2Tzt10GFK2Ro9tAVEtacMqRmTarYMFKAnV2vYMdwWc73xzmDQyAV7SwEdMhzmiRoo7+m88DuiXlJlGeaw==","cpu": ["arm64"],@@ -596,9 +596,9 @@}},"node_modules/@next/swc-darwin-x64": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/swc-darwin-x64/-/swc-darwin-x64-15.5.13.tgz",- "integrity": "sha512-Ey3fuUeWDWtVdgiLHajk2aJ74Y8EWLeqvfwlkB5RvWsN7F1caQ6TjifsQzrAcOuNSnogGvFNYzjQlu7tu0kyWg==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/swc-darwin-x64/-/swc-darwin-x64-15.5.15.tgz",+ "integrity": "sha512-G+YNV+z6FDZTp/+IdGyIMFqalBTaQSnvAA+X/hrt+eaTRFSznRMz9K7rTmzvM6tDmKegNtyzgufZW0HwVzEqaQ==","cpu": ["x64"],@@ -612,9 +612,9 @@}},"node_modules/@next/swc-linux-arm64-gnu": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/swc-linux-arm64-gnu/-/swc-linux-arm64-gnu-15.5.13.tgz",- "integrity": "sha512-aLtu/WxDeL3188qx3zyB3+iw8nAB9F+2Mhyz9nNZpzsREc2t8jQTuiWY4+mtOgWp1d+/Q4eXuy9m3dwh3n1IyQ==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/swc-linux-arm64-gnu/-/swc-linux-arm64-gnu-15.5.15.tgz",+ "integrity": "sha512-eVkrMcVIBqGfXB+QUC7jjZ94Z6uX/dNStbQFabewAnk13Uy18Igd1YZ/GtPRzdhtm7QwC0e6o7zOQecul4iC1w==","cpu": ["arm64"],@@ -628,9 +628,9 @@}},"node_modules/@next/swc-linux-arm64-musl": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/swc-linux-arm64-musl/-/swc-linux-arm64-musl-15.5.13.tgz",- "integrity": "sha512-9VZ0OsVx9PEL72W50QD15iwSCF3GD/dwj42knfF5C4aiBPXr95etGIOGhb8rU7kpnzZuPNL81CY4vIyUKa2xvg==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/swc-linux-arm64-musl/-/swc-linux-arm64-musl-15.5.15.tgz",+ "integrity": "sha512-RwSHKMQ7InLy5GfkY2/n5PcFycKA08qI1VST78n09nN36nUPqCvGSMiLXlfUmzmpQpF6XeBYP2KRWHi0UW3uNg==","cpu": ["arm64"],@@ -644,9 +644,9 @@}},"node_modules/@next/swc-linux-x64-gnu": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/swc-linux-x64-gnu/-/swc-linux-x64-gnu-15.5.13.tgz",- "integrity": "sha512-3knsu9H33e99ZfiWh0Bb04ymEO7YIiopOpXKX89ZZ/ER0iyfV1YLoJFxJJQNUD7OR8O7D7eiLI/TXPryPGv3+A==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/swc-linux-x64-gnu/-/swc-linux-x64-gnu-15.5.15.tgz",+ "integrity": "sha512-nplqvY86LakS+eeiuWsNWvfmK8pFcOEW7ZtVRt4QH70lL+0x6LG/m1OpJ/tvrbwjmR8HH9/fH2jzW1GlL03TIg==","cpu": ["x64"],@@ -660,9 +660,9 @@}},"node_modules/@next/swc-linux-x64-musl": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/swc-linux-x64-musl/-/swc-linux-x64-musl-15.5.13.tgz",- "integrity": "sha512-AVPb6+QZ0pPanJFc1hpx81I5tTiBF4VITw5+PMaR1CrboAUUxtxn3IsV0h48xI7fzd6/zw9D9i6khRwME5NKUw==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/swc-linux-x64-musl/-/swc-linux-x64-musl-15.5.15.tgz",+ "integrity": "sha512-eAgl9NKQ84/sww0v81DQINl/vL2IBxD7sMybd0cWRw6wqgouVI53brVRBrggqBRP/NWeIAE1dm5cbKYoiMlqDQ==","cpu": ["x64"],@@ -676,9 +676,9 @@}},"node_modules/@next/swc-win32-arm64-msvc": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/swc-win32-arm64-msvc/-/swc-win32-arm64-msvc-15.5.13.tgz",- "integrity": "sha512-FZ/HXuTxn+e5Lp6oRZMvHaMJx22gAySveJdJE0//91Nb9rMuh2ftgKlEwBFJxhkw5kAF/yIXz3iBf0tvDXRmCA==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/swc-win32-arm64-msvc/-/swc-win32-arm64-msvc-15.5.15.tgz",+ "integrity": "sha512-GJVZC86lzSquh0MtvZT+L7G8+jMnJcldloOjA8Kf3wXvBrvb6OGe2MzPuALxFshSm/IpwUtD2mIoof39ymf52A==","cpu": ["arm64"],@@ -692,9 +692,9 @@}},"node_modules/@next/swc-win32-x64-msvc": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/@next/swc-win32-x64-msvc/-/swc-win32-x64-msvc-15.5.13.tgz",- "integrity": "sha512-B5E82pX3VXu6Ib5mDuZEqGwT8asocZe3OMMnaM+Yfs0TRlmSQCBQUUXR9BkXQeGVboOWS1pTsRkS9wzFd8PABw==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/@next/swc-win32-x64-msvc/-/swc-win32-x64-msvc-15.5.15.tgz",+ "integrity": "sha512-nFucjVdwlFqxh/JG3hWSJ4p8+YJV7Ii8aPDuBQULB6DzUF4UNZETXLfEUk+oI2zEznWWULPt7MeuTE6xtK1HSA==","cpu": ["x64"],@@ -1607,9 +1607,9 @@}},"node_modules/lodash": {- "version": "4.17.23",- "resolved": "https://registry.npmjs.org/lodash/-/lodash-4.17.23.tgz",- "integrity": "sha512-LgVTMpQtIopCi79SJeDiP0TfWi5CNEc/L/aRdTh3yIvmZXTnheWpKjSZhnvMl8iXbC1tFg9gdHHDMLoV7CnG+w==",+ "version": "4.18.1",+ "resolved": "https://registry.npmjs.org/lodash/-/lodash-4.18.1.tgz",+ "integrity": "sha512-dMInicTPVE8d1e5otfwmmjlxkZoUpiVLwyeTdUsi/Caj/gfzzblBcCE5sRHV/AsjuCmxWrte2TNGSYuCeCq+0Q==","license": "MIT"},"node_modules/loose-envify": {@@ -1662,12 +1662,12 @@}},"node_modules/next": {- "version": "15.5.13",- "resolved": "https://registry.npmjs.org/next/-/next-15.5.13.tgz",- "integrity": "sha512-n0AXf6vlTwGuM93Z++POtjMsRuQ9pT5v2URPciXKUQIl/EB2WjXF0YiIUxaa9AEMFaMpZlaG3KPK6i4UVnx9eQ==",+ "version": "15.5.15",+ "resolved": "https://registry.npmjs.org/next/-/next-15.5.15.tgz",+ "integrity": "sha512-VSqCrJwtLVGwAVE0Sb/yikrQfkwkZW9p+lL/J4+xe+G3ZA+QnWPqgcfH1tDUEuk9y+pthzzVFp4L/U8JerMfMQ==","license": "MIT","dependencies": {- "@next/env": "15.5.13",+ "@next/env": "15.5.15","@swc/helpers": "0.5.15","caniuse-lite": "^1.0.30001579","postcss": "8.4.31",@@ -1680,14 +1680,14 @@"node": "^18.18.0 || ^19.8.0 || >= 20.0.0"},"optionalDependencies": {- "@next/swc-darwin-arm64": "15.5.13",- "@next/swc-darwin-x64": "15.5.13",- "@next/swc-linux-arm64-gnu": "15.5.13",- "@next/swc-linux-arm64-musl": "15.5.13",- "@next/swc-linux-x64-gnu": "15.5.13",- "@next/swc-linux-x64-musl": "15.5.13",- "@next/swc-win32-arm64-msvc": "15.5.13",- "@next/swc-win32-x64-msvc": "15.5.13",+ "@next/swc-darwin-arm64": "15.5.15",+ "@next/swc-darwin-x64": "15.5.15",+ "@next/swc-linux-arm64-gnu": "15.5.15",+ "@next/swc-linux-arm64-musl": "15.5.15",+ "@next/swc-linux-x64-gnu": "15.5.15",+ "@next/swc-linux-x64-musl": "15.5.15",+ "@next/swc-win32-arm64-msvc": "15.5.15",+ "@next/swc-win32-x64-msvc": "15.5.15","sharp": "^0.34.3"},"peerDependencies": {diff --git a/awesome_agent_skills/self-improving-agent-skills/frontend/src/components/RunningStep.tsx b/awesome_agent_skills/self-improving-agent-skills/frontend/src/components/RunningStep.tsxindex 8554343a..88ff6c25 100644--- a/awesome_agent_skills/self-improving-agent-skills/frontend/src/components/RunningStep.tsx+++ b/awesome_agent_skills/self-improving-agent-skills/frontend/src/components/RunningStep.tsx@@ -44,7 +44,7 @@ export default function RunningStep({const startOptimization = async () => {setIsRunning(true);- const API_BASE = process.env.NEXT_PUBLIC_API_URL || "http://localhost:8892";+ const API_BASE = process.env.NEXT_PUBLIC_API_URL || "http://localhost:8891";try {// Start the optimization