Context, not clutter
Find the smallest useful capability for the task. Compose atomic Skills into repeatable delivery systems.
Browse artifacts ↗ THE TRUST LAYER FOR AI DEVELOPMENT
Build with
signal.
A living library of Skills, Workflows and Playbooks that helps your agents choose the right way to work — with provenance, context and confidence.
Open discovery · Agent-ready · Source-linked
LIVE REGISTRY● indexed
—
artifacts ready for your agent
safe-feature-deliveryorchestrator · v8
source verifiedimmutable provenance
SKILLS✦WORKFLOWS✦PLAYBOOKS✦POLICIES✦EVALUATIONS✦BEST PRACTICES✦ORCHESTRATORS✦
01 / WHY SKILLHUB
Less guessing.
More shipping.
Agents are only as good as the decisions around their work. SkillHub gives them a map: what to use, why it fits, where it came from and what to do next.
Trust you can inspect
Every result carries source, version, license, risk signals and related artifacts so your agent can show its work.
See the signals ↗Ready in your workflow
Connect once through MCP. Let agents search, compare and recommend while installation stays explicit and human-approved.
Connect MCP ↗02 / THE LIBRARY
One registry.
Every layer.
Search the shared operating system for agent-assisted development. Start with a task, then follow the connections.
03 / AGENT CONNECTION
Your agent,
in the loop.
Create an account, generate a personal MCP token and paste the ready-made config into Claude Code or another MCP client.
{
"mcpServers": {
"maidb": {
"url": "https://mcp.maidb.dev/mcp",
"headers": { "Authorization": "Bearer sh_••••••••" }
}
}
} personal token · read-only discovery