{"slug": "stop-your-ai-agent-from-building-code-on-abandoned-2022-libraries-open-source", "title": "Stop Your AI Agent From Building Code on Abandoned 2022 Libraries (Open-Source MCP Server)", "summary": "A developer released skills-but-for-dev, an open-source MCP server that acts as a pre-flight curator for AI coding agents such as Cursor, Claude Code and Copilot, checking live GitHub activity, PyPI download counts and Hugging Face VRAM requirements before the agent commits to a library. The server scores candidates with an 80% base, 10% health and 10% PyPI weighting, filters out GPL and non-commercial licenses, and caches results in SQLite for roughly 4.7ms warm latency. It runs via uvx with no installation and auto-detects GitHub credentials through gh auth token.", "body_md": "**Intelligent technology stack and open-source curator for AI coding agents.**\n\n*Never let your AI agent build production code on abandoned 2022 libraries.*\n\nWhen you ask Cursor, Claude Code, Antigravity, or Copilot:\n\n*\"Build an AI meeting assistant that records audio, separates speakers, and generates summaries\"*\n\nThe model immediately starts writing code using libraries from its **static pre-training data**. Meanwhile:\n\n`GPL-3.0` or `CC-BY-NC-4.0` libraries can create major intellectual property and commercial liabilities.\n**`skills-but-for-dev` fixes this.** It acts as an autonomous pre-flight curator for your agent, querying live GitHub velocity, PyPI monthly adoption, and Hugging Face VRAM requirements **before** writing code.\n\n| Task: *\"Build a local meeting notes agent\"* | ❌ Vanilla AI Agent (Without `skills-but-for-dev` ) | ✅ AI Agent + `skills-but-for-dev` | \n|---|---|---|\n| **Transcription (ASR)** | Suggests `SpeechRecognition` (deprecated Python 2 era wrapper) | **`Systran/faster-whisper`** (9.6M/mo PyPI downloads, 4x faster CTranslate2) | \n| **Speaker Diarization** | Tries to hand-code clustering or hallucinates non-existent APIs | **`modelscope/FunASR`** /**` pyannote.audio`** (Verified active acceleration: $v_4 / v_{12} = 1.2$) | \n| **Licensing** | Generates code importing `GPL` / non-commercial packages | **100% Permissive Verification** (Flags MIT / Apache-2.0, filters restrictive licenses) | \n| **Local Hardware Check** | Recommends large Whisper model, crashing local GPU | **Automatic VRAM Baseline** (Estimates 0.54 GB FP16 ceiling for`faster-whisper-small` ) | \n\n`stars:>500`) alongside fast-rising breakout repositories (` stars:75..3500`, pushed $< 30$ days) without cannibalization.` pypistats.org` to protect against star-inflation and GitHub bots.`.safetensors`, sharded model weights, Diffusers pipelines, and distinct GGUF quantization levels (`Q4_K_M`, `Q8_0`).` gh auth token`) or falls back to an adaptive, quota-conserving unauthenticated mode.\n\n```\nflowchart TD\n    User[\"Developer Prompt\\n'Build a RAG system for PDF invoices'\"] --> Agent[\"AI Coding Agent\\n(Cursor / Antigravity / Claude Code)\"]\n    Skill[\"SKILL.md\"] -.->|\"Decomposition Instructions\"| Agent\n    Agent -->|\"1. search_github_repos('pdf parser')\\n2. search_hf_models('embeddings')\"| Server[\"skills-but-for-dev MCP Server\"]\n\n    subgraph Server_Internals[\"skills-but-for-dev Engine\"]\n        Server --> Norm[\"Query Normalizer\\n(Noise stripping & Language filter)\"]\n        Norm --> GH[\"GitHub REST Client\\n(Dual-Bucket search + Star Velocity)\"]\n        Norm --> HF[\"Hugging Face Client\\n(VRAM & Pipeline Estimator)\"]\n        Norm --> PyPI[\"PyPI Stats Client\\n(30-Day Download Verification)\"]\n        GH --> Scoring[\"Conservative Scoring Engine\\n(80% Base + 10% Health + 10% PyPI)\"]\n        HF --> Scoring\n        PyPI --> Scoring\n        Scoring --> Cache[(\"SQLite Cache (WAL Mode)\\n~4.7ms warm latency\")]\n    end\n\n    Server -->|\"Curated Architecture Recipe\"| Agent\n    Agent -->|\"Step-by-step Tech Stack with Verified Licences\"| User\n```\n\n`uvx` (Zero Installation)\nNo cloning or virtual environment setup required. Works with any MCP-compatible client:\n\n```\n# Verify it runs and check diagnostics\nuvx skills-but-for-dev ping\n```\n\n`.cursor/mcp.json`)\n\n```\n{\n  \"mcpServers\": {\n    \"skills-but-for-dev\": {\n      \"command\": \"uvx\",\n      \"args\": [\"skills-but-for-dev\"],\n      \"env\": {\n        \"GITHUB_TOKEN\": \"optional_your_pat_here\"\n      }\n    }\n  }\n}\n```\n\n`mcp_config.json`)\n\n```\n{\n  \"mcpServers\": {\n    \"skills-but-for-dev\": {\n      \"command\": \"uvx\",\n      \"args\": [\"skills-but-for-dev\"]\n    }\n  }\n}\n```\n\n`claude_desktop_config.json`)\n\n```\n{\n  \"mcpServers\": {\n    \"skills-but-for-dev\": {\n      \"command\": \"uvx\",\n      \"args\": [\"skills-but-for-dev\"]\n    }\n  }\n}\n```\n\n**Tip:** If you already use the GitHub CLI (`gh`), `skills-but-for-dev` automatically detects your authentication token via `gh auth token`. You don't even need to configure `GITHUB_TOKEN` manually!\n\nWhen your agent uses `skills-but-for-dev`, it generates structured architecture recipes before writing code:\n\n| Layer | Recommended Component | Star Velocity | Real-World PyPI Adoption | License | Hardware / Memory | \n|---|---|---|---|---|---|\n| **ASR (Speech-to-Text)** | [`SYSTRAN/faster-whisper`](https://github.com/SYSTRAN/faster-whisper) | $v_4 = 18.2$/day (accel: 1.15) | **9,614,222** downloads/mo | MIT | ~0.54 GB VRAM (small FP16) | \n| **Speaker Diarization** | [`modelscope/FunASR`](https://github.com/modelscope/FunASR) | $v_4 = 8.4$/day (accel: 1.20) | **272,534** downloads/mo | MIT | ~1.2 GB VRAM | \n| **Embeddings** | [`sentence-transformers`](https://github.com/UKPLab/sentence-transformers) | $v_4 = 14.1$/day (accel: 1.05) | **22,332,616** downloads/mo | Apache-2.0 | ~0.62 GB VRAM | \n\nBenchmarked across realistic multi-layer agent scenarios using strict Ground Truth matching:\n\n| Metric | Result | Benchmark Details | \n|---|---|---|\n| **Warm Cache Latency** | **4.73 ms** | Sub-5 millisecond response time via SQLite WAL mode | \n| **Cold API Latency** | **1,728 ms** | Optimized single-flight queries with adaptive quotas | \n| **Cache Speedup** | **365.7x** | Massive latency reduction for recurring architecture queries | \n| **Commercial License Safety** | **95.2%** | Automatic filtering of GPL / CC-BY-NC restrictive licenses | \n| **PyPI Download Signal Rate** | **52.4%** | Over half of returned candidates enriched with production downloads | \n| **Unit Test Coverage** | **40 / 40 Passing** | Comprehensive suites for normalizer, PyPI client, cache, and query relaxation | \n\nYou can run the evaluation suite locally at any time:\n\n```\npython evals/run_evals.py\n```\n\nInspect active tokens, cache paths, and server health directly from your terminal:\n\n```\n# Print active version\nskills-but-for-dev --version\n\n# Ping server and verify token discovery source (ENV vs GitHub CLI)\nskills-but-for-dev ping\n```\n\nFor mathematical formulas (star velocity moving averages, exponential recency decay, dormant spike detection ratios), see [ARCHITECTURE.md](//ARCHITECTURE.md).\n\nDistributed under the [Apache License 2.0](https://dev.toLICENSE).", "url": "https://wpnews.pro/news/stop-your-ai-agent-from-building-code-on-abandoned-2022-libraries-open-source", "canonical_source": "https://dev.to/melihcan_zdemir_ae6aa90b/stop-your-ai-agent-from-building-code-on-abandoned-2022-libraries-open-source-mcp-server-2c3n", "published_at": "2026-10-28 14:34:05+00:00", "updated_at": "2026-10-28 14:41:31.037059+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "developer-tools", "agent-protocols", "ai-products"], "entities": ["skills-but-for-dev", "Cursor", "Claude Code", "GitHub", "Hugging Face", "PyPI", "Copilot", "Antigravity"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/stop-your-ai-agent-from-building-code-on-abandoned-2022-libraries-open-source", "markdown": "https://wpnews.pro/news/stop-your-ai-agent-from-building-code-on-abandoned-2022-libraries-open-source.md", "text": "https://wpnews.pro/news/stop-your-ai-agent-from-building-code-on-abandoned-2022-libraries-open-source.txt", "jsonld": "https://wpnews.pro/news/stop-your-ai-agent-from-building-code-on-abandoned-2022-libraries-open-source.jsonld"}}