Every developer knows this feeling.
You need a vector database. Or a job queue. Or a .dwg parser for Node.js. You open a browser tab, search GitHub, open five more tabs, try to compare stars and last-commit dates, get distracted by a Medium post from 2019, and 45 minutes later you've picked something based on vibes.
There's a better way. I built it.
SKILLmama is an AI-native capability discovery engine. Point it at a capability gap, and it scans your project architecture, searches across five tiers of the ecosystem, scores every candidate against your specific stack, and returns ranked recommendations with evidence and clickable links.
It works with Claude Code, Claude.ai, OpenAI Codex, and Antigravity.
Most library selection looks like this:
The real question isn't "what's popular" — it's "what fits my stack, has active maintenance, and won't take a week to integrate?" Those are four separate signals and you need all four weighted correctly.
SKILLmama scores every candidate on four dimensions:
Score = (Compatibility × 0.40) +
(Popularity × 0.30) +
(Maintenance × 0.15) +
(Simplicity × 0.15)
| Factor | Weight | What it measures |
|---|---|---|
| Compatibility | 40% | Language/framework fit, official SDK, integration effort |
| Popularity | 30% | GitHub stars, npm/PyPI weekly downloads |
| Maintenance | 15% | Days since last commit, release cadence |
| Simplicity | 15% | Setup effort, documentation quality |
Compatibility is weighted highest because a library with 50k stars that doesn't have a Python client is useless if your stack is Python. Popularity comes second because the ecosystem around a library matters. Maintenance and simplicity round it out.
Every score is 1–10 per dimension. No black box. You can see exactly why something ranked #1.
SKILLmama doesn't just search one place. It works through five tiers in order, stopping when it has 8+ candidates:
| Tier | Source | What it finds |
|---|---|---|
| 1 | skills.sh | Reusable skills and capability patterns |
| 2 | GitHub | Open-source libraries, frameworks, SDKs |
| 3 | Smithery / MCP Ecosystem | AI-native tools installable as MCP servers |
| 4 | npm / PyPI / pkg.go.dev | Package registries with download signals |
| 5 | Curated Templates | LangGraph, OpenHands, cookbook examples |
Tier 3 is the interesting one. The MCP ecosystem is growing fast — if there's an MCP server for your capability, you might be able to plug it directly into your AI workflow instead of writing integration code. SKILLmama surfaces that option explicitly.
┌─────────────────────────────────────────────────────────┐
│ USER REQUEST │
└─────────────────────────┬───────────────────────────────┘
│
▼
┌───────────────────────┐
│ PHASE 0 │
│ Understand Request │
│ Extract: capability,│
│ stack, constraints │
└───────────┬───────────┘
│
▼
◇ Capability
vague?
/ \
YES NO
│ │
▼ │
Ask 1 clarifying │
question, await │
user response │
│ │
└──────┬──────┘
│
▼
┌───────────────────────┐
│ PHASE 1 │
│ Architecture Scan │
└───────────┬───────────┘
│
▼
◇ In a project
repo?
/ \
YES NO
│ │
▼ │
Read: package.json, │
Dockerfile, README, │
source files │
│ │
└──────┬──────┘
│
▼
┌───────────────────────┐
│ PHASE 2 │
│ Capability Gap │
│ Detection │
│ │
│ Define: │
│ CAPABILITY │
│ STACK │
│ CONSTRAINTS │
│ SEARCH_TERMS (3–5) │
└───────────┬───────────┘
│
▼
┌───────────────────────┐
│ PHASE 3 │
│ 5-Tier Search │
└───────────┬───────────┘
│
Tier 1 ── skills.sh
↓
Tier 2 ── GitHub (stars, recency, contrib)
↓
Tier 3 ── MCP Ecosystem
↓
Tier 4 ── npm / PyPI registries
↓
Tier 5 ── Templates & Cookbooks
│
◇ 8+ candidates found?
/ \
YES NO
│ │
Skip remaining Continue tiers
tiers │
└──────┬─────────┘
│
▼
┌────────────────────────────────────────┐
│ PHASE 4 — Score Each Candidate │
│ │
│ Score = (C × 0.40) + │
│ (P × 0.30) + │
│ (M × 0.15) + │
│ (S × 0.15) │
└───────────────┬────────────────────────┘
│
▼
┌────────────────────────────────────────┐
│ PHASE 5 — Present Results │
│ │
│ #1, #2, #3 — full score breakdown │
│ Also Considered — table │
│ MCP callout (if found) │
│ Next Steps (3 actions) │
└────────────────────────────────────────┘
Prompt:
/skillmama find me a vector database for my FastAPI + Python project
Phase 1 — Architecture Scan
SKILLmama reads your project files first:
✓ pyproject.toml → Python 3.11, FastAPI, SQLAlchemy
✓ Dockerfile → containerized, no GPU
✓ .env.example → OPENAI_API_KEY present → RAG use case confirmed
Detected stack: Python / FastAPI / PostgreSQL / Docker / OpenAI
Phase 2 — 5-Tier Search
Tier 1 skills.sh → "qdrant-memory-skill", "chroma-rag-skill"
Tier 2 GitHub → qdrant (17k★), chroma (14k★), pgvector (11k★), milvus (29k★)
Tier 3 MCP → qdrant-mcp-server, chroma-mcp
Tier 4 PyPI → qdrant-client (380k/wk), chromadb (620k/wk), pgvector (180k/wk)
Tier 5 Templates → LangChain + Qdrant RAG template, FastAPI + Chroma starter
Phase 3 — Scoring
| Candidate | Compat | Popular | Maint | Simple | Score |
|---|---|---|---|---|---|
| Qdrant | 9 | 8 | 10 | 9 | 9.05 |
| pgvector | 10 | 7 | 9 | 7 | 8.65 |
| Chroma | 8 | 8 | 9 | 10 | 8.55 |
| Milvus | 6 | 9 | 10 | 4 | 7.05 |
Phase 4 — Output
#1 — Qdrant · Score: 9.05/10
Docker-native, official Python client, active MCP server for AI workflows.
- Compatibility: 9/10 — official qdrant-client SDK, FastAPI examples in docs
- Popularity: 8/10 — 17k GitHub stars, 380k PyPI downloads/week
- Maintenance: 10/10 — committed 2 days ago, weekly releases
- Simplicity: 9/10 — docker run + pip install, 5-min setup
- Install: docker run -p 6333:6333 qdrant/qdrant && pip install qdrant-client
#2 — pgvector · Score: 8.65/10
Stay on PostgreSQL — no new infra, native SQL queries.
#3 — Chroma · Score: 8.55/10
Easiest local dev setup; best for prototyping before scaling.
MCP Option: qdrant-mcp-server — install as MCP tool for direct AI memory integration.
Next Steps:
1. docker run qdrant/qdrant and pip install qdrant-client to validate locally
2. Use the LangChain + Qdrant RAG template as a starting point
3. If staying Postgres-only, evaluate pgvector — saves an infra hop
45 minutes of tab-hopping, compressed into a structured decision.
Claude Code:
mkdir -p /your-project/.claude/commands
cp .claude/commands/skillmama.md /your-project/.claude/commands/skillmama.md
Then type /skillmama
in any Claude Code session.
Claude.ai:
zip -r skillmama.zip skillmama/
/skillmama
in any conversationOpenAI Codex:
Place codex/AGENTS.md
in your repo root, then ask naturally:
codex "find me the best job queue for this project"
Antigravity:
Load antigravity/PROMPT.md
as the system prompt, then ask naturally.
All four adapters run the same pipeline and produce the same output format.
Not an IDE. Not autocomplete. Not a chatbot.
It's a capability oracle — it tells you what to use and why, with evidence. You still write the code. SKILLmama just makes sure you're writing it with the right tool.
Apache 2.0 licensed. Works with Claude Code, Claude.ai, OpenAI Codex, and Antigravity. A pre-built skillmama.zip
is included for Claude.ai upload — no build step needed.
/skillmama find me a vector database for my FastAPI project
/skillmama what auth library should I use for my Next.js app?
/skillmama scan my project and tell me what's missing
/skillmama find a .dwg parser for Node.js
What's the most time you've spent picking a library, only to switch it out later? Drop it in the comments.