The trust layer between AI and your database, so an agent's answers are defensible. Every join, metric, and named filter is confidence-scored and either FK-derived or human-approved via a review dashboard. Every answer ships a SQL receipt showing what ran and which model version it used. Works across Postgres / MySQL…
Earth data, agent-ready: a Development Seed labs prototype. MCP tools over the cloud-native geospatial stack plus the earth-data skill and a groundstation skill that runs and demos it.
★not rated 0
changed yesterdayA
tokens not measured
originalMIT
Connect Claude Code to any ObjectStack deployment over MCP: discover objects, query and change records, and run business actions — under your own permissions and row-level security. OAuth self-serve; the deployment URL is the only input.
★not rated 0 4mo agoA
tokens not measured
originalApache-2.0
Long-video vision for coding agents — coarse-to-fine map/zoom navigation, storyboard montages, and an OCR sidecar for on-screen code. External-free: the agent's own model does the watching.
★not rated 0 4mo agoA
tokens not measured
originalMIT
Stops AI coding assistants from using things that don't exist — blocks hallucinated/slop-squatted package installs and dead file paths (fail-closed hook + verifyidentifier MCP tool).
★not rated 0 4mo agoA
tokens not measured
originalApache-2.0
ideate — a composable process framework for Claude Code: an append-only process record, a dependency-ordered work board, and gated steering, with skills, agents, hooks, and CLIs that drive an autopilot loop from idea to verified work.
★not rated 0
changed 1mo agoA
tokens not measured
AGPL-3.0
Connect Claude Code to a live DASE ORM Designer session in VS Code: read, edit and organize .dsorm data models (tables, fields, FK references, layout) through 40+ MCP tools.
★not rated 0 3mo agoA
tokens not measured
originalMIT
Validate UBL 2.1 invoices against the official EN 16931 Schematron and generate conformant EU e-invoices, directly from your agent. No install, no API key — the catalog is public and every tool is safe to call.
★not rated 0 3mo agoA
tokens not measured
originalMIT
Connect coding agents to Alabasta tasks: read tasks (MCP), a skill that teaches the resolution protocol, and a Stop hook that enforces a structured resolution before finishing.