mcp · Global

geoai-skills

Measured Agent Skills for geospatial AI — 18 skills that stop when the data cannot support the claim. Routing benchmark, held-out split, and known defects published.

Visit source ↗GitHub repository ↗

Claim this listing

Represent geoai-skills? Send a claim for review so we can discuss listing updates. Claims do not automatically grant ownership or verification.

Explore paid sponsorship →

What's in it

Data offered docs-verified: Agent Skills that set preconditions, required checks and evidence reporting for geospatial analysis claims · Stage coverage: route and acquire, sense and prepare, model and detect, and related method guardrails
Coverage llm-inferred: Global
Formats docs-verified: Agent Skills (SKILL.md)
Access docs-verified: Free and open; MIT licence. No credentials documented. Installed via npx skills add.

Useful for

Use cases llm-inferred: Check whether two Sentinel-2 scenes from different seasons can support a changed-hectares claim before any pixel arithmetic · Verify CRS and units before reporting flood extent or catchment area figures · Audit spatial leakage before reporting validation scores for a flood or susceptibility model · Confirm DEM and terrain-hydrology preconditions before reporting flow-path outputs
Audience llm-inferred: GIS analysts · remote sensing analysts · hydrologists · flood consultants · geospatial developers building AI agent workflows

Limitations

Constraints docs-verified: Behaviour figures come from one runtime and model pair (Claude Code 2.1.214/2.1.222, claude-sonnet-5) and are not universal · Behaviour figures are model-judged, uncalibrated and not yet human-verified · The repository's own Phase 3 quality gate fails: 4.3% critical failures and 19.6% full-criterion pass rate · Known routing defects: 1 false positive, 4 false negatives, 1 incomplete route and 11 execution errors · Routing shows which skill loaded, not whether the answer was correct · Contains no data feeds or UK-specific datasets; it supplies method guidance only · ENTRY kind is 'mcp' and its tools list names skills, but the docs describe Agent Skills rather than an MCP server

Licence: MIT
Authentication: No credentials documented in README
GitHub stars: 29
Last commit: 14/10/2026
Last successful page check: 14/10/2026
Page health: healthy
Package: npm:skills
Verification: Not independently verified

Provenance & evidence

Geospatial deep dive (Glama/Smithery/GitHub). Tools/install/auth auto-extracted from README.

Metadata observed 15/10/2026 · research sweep: geo-deep-dive (geo-deep-dive:github-search-geo)