geoai-skills
About
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.
Details
- Author
- muend
- Categories
- Productivity
Jump to
Setup
Install geoai-skills in your MCP client (Claude Desktop, Cursor, Windsurf, and others).
Repository: https://github.com/muend/geoai-skills
Follow the installation instructions in the repository README, then restart your MCP client.
The correctness layer for geospatial AI agents
Stop silent CRS, spatial-leakage, validity, unit, and uncertainty failures before they ship.
GeoAI Skills is a vendor-neutral collection of 18Agent Skillsthat turns a general-purpose AI agent into a more defensible geospatial collaborator. It covers the full workflow—from STAC search and PostGIS to spatial statistics, Earth observation, LiDAR, cartography, and guarded ArcGIS Pro automation—while making verification and limitations part of the deliverable.
These figures describe the shipped tree, on one runtime/model pair— Claude Code2.1.214withclaude-sonnet-5, 167-case suiteefe27d8c1736…, 2026-08-05 — and are not universal or model-independent claims. Routing measures which skill activates, not whether the answer is correct;answer quality has never been measured.The run that produced these numbers also recorded one false positive and two cases that moved from pass to fail without an established cause. SeeBENCHMARK.mdfor the full card, the reviewed failures, and what the run does not show.
A plausible shortcut is not a defensible result.
Watch the short MP4·View the static verdict·Install·Inspect the evidence
The demo uses a public, synthetic prompt. It shows the intended guardrail behavior, not a claim that every model or runtime will produce identical wording.
npx skills add muend/geoai-skills --skill '' -a codex
Or install one specialist for Claude Code:
npx skills add muend/geoai-skills \ --skill remote-sensing-analysis \ -a claude-code
Can May 2024 and September 2025 Sentinel-2 scenes support a defensible changed-hectares claim?
The relevant skills should activate automatically. Instead of blindly subtracting two rasters, the agent is instructed to test comparability, identify the phenological mismatch, withhold an unsupported area claim, and specify the evidence needed to proceed.
What changes when the skills are present
These skills complement MCP servers, GIS libraries, and hosted tools. They are the method and verification layer that tells an agentwhen not to trust an apparently successful operation.
- a buffer computed in degrees still returns numbers;
- a random spatial train/test split still produces a beautiful learning curve;
- a misregistered change map still shows crisp-looking boundaries;
- overlapping polygons can silently inflate an area total;
- an attractive choropleth can still encode the wrong class semantics.
General-purpose models often know the APIs. The harder problem is knowing which preconditions, controls, and refusal conditions make a geospatial claim defensible. GeoAI Skills encodes that discipline:
- CRS and units are explicit.
- Spatial leakage is treated as a default risk.
- Every stage ends with numeric and visual verification.
- Uncertainty and sensitivity are outputs, not optional footnotes.
- Missing evidence narrows or blocks the claim instead of being guessed.
The current source tree contains two deliberately separate evaluation layers:
The external subset is independently authored and reported separately. It does not copy upstream datasets, prompts, or reference implementations, and its outcomes must not be pooled with native routing metrics. Its five-case v1 source population is frozen under suite hashc99563100cac…; a separate v2 contract freeze distinguishes semantic correctness, evidence sufficiency, and exact representation compliance without changing those cases. Frozen producer interface v1 preserves the first shape-only condition; v2 adds answer-safe method, serialization, UTF-8, and workspace execution guidance after observed cross-runtime ambiguities, without exposing reference values. The separately frozen v3 condition pairs a semantic validator with fully disclosed nested output structures, bounded floating-point tolerance, exact inventory checks, and representation-independent map checks; historical v1/v2 evidence remains unchanged. Every deterministically rendered prompt is hashed. Results follow theoffline run and result protocol, which reports skill activation, runtime completion, artifact-contract compliance, and overall success as separate measures.
An 18-skill, 167-case Claude Code run recorded99.17% routing precision, 96.77% routing recall, and 96.41% full-route accuracyon suiteefe27d8c1736…under Claude Code2.1.214withclaude-sonnet-5(2026-08-05). The paired skills-disabled control produced zero activations across all 167 cases. All four false negatives, the one false positive, and all elevenmax_turnsexecution errors were inspected; every enabled error had its activation recovered from the raw trace, none lost, so the enabled routing metrics are unaffected. Behavior quality remainsunclaimeduntil independent-family judging and the manual-review protocol are complete.
This run was bought to test a published defect, and it also found new ones.The previous card attributed four of its nine false negatives to one clause in thechange-detectiondescription and shipped the fix as a hypothesis. All four of those cases now route correctly, both over-correction guards held, and the control arm is clean. Against that: one boundary probe still misroutes, and two cases that previously passed now fail without a cause the diff can explain — run-to-run variance and a real regression are indistinguishable from a single run. The suites are also different populations (158 → 167 cases), so the figures above arenota before/after ratio; the honest comparison is case-level and is in the card.
ReadBENCHMARK.mdfor the full result card andEVALUATION.mdfor the provider-neutral protocol, suite hashes, split rules, judge boundaries, and publication gates.
Install the full suite for cross-skill routing, or cherry-pick a specialist. Every skill must remain safe and useful alone; sibling references are advisory and critical safeguards have local fallbacks.
Browse the collection onskills.shor inspect it without installing:
npx skills add muend/geoai-skills --list
npx skills add muend/geoai-skills --skill '' -a claude-code
npx skills add muend/geoai-skills --skill '' -a codex
Runnpx skills add muend/geoai-skillswithout flags for the interactive agent and skill picker.
Version0.2.0is published in the OpenAI Plugins Directory asGeoAI Skills. The repository is a skills-only plugin: it adds no hosted service, authentication flow, or MCP server.
python tools/build_openai_plugin_bundle.py
The ignoreddist/output contains only the plugin manifest, public policy pages, logo, and runtime skill files. Evaluation cases, benchmark traces, repository automation, and private development material are excluded.
claude plugin marketplace add muend/geoai-skills claude plugin install geoai@geoai-skills
From an interactive Claude Code session:
/plugin marketplace add muend/geoai-skills /plugin install geoai@geoai-skills
Installed skills use the stablegeoainamespace, for example:
GitHub Copilot can load project skills from.github/skills,.claude/skills, or.agents/skills. With GitHub CLI 2.90.0 or later, preview before installing:
gh skill preview muend/geoai-skills remote-sensing-analysis gh skill install muend/geoai-skills remote-sensing-analysis
The Skills CLI also provides an explicit Copilot target:
npx skills add muend/geoai-skills \ --skill remote-sensing-analysis \ -a github-copilot
See GitHub'sAgent Skills documentationfor supported locations, pinning, provenance, and security guidance.
For Claude.ai or Claude desktop, download individual skill ZIP files andSHA256SUMSfrom the latest GitHub Release, then upload them throughCustomize → Skills → Create skill → Upload a skill*.
To build the same deterministic archives locally:
Any Agent-Skills-compatible runtime can instead copy directories fromskills/into its skills directory. Realarcgis-pro-automationexecution additionally requires Windows, licensed ArcGIS Pro, and a configured localarcgis-mcp-bridge.
Release maintainers should useRELEASING.mdfor the non-publishing candidate workflow, clean-install matrix, evidence requirements, publication gate, and rollback procedure.
Plan a defensible two-date Sentinel-2 forest-loss workflow. One scene is from May 2024 and the other from September 2025. Decide whether this comparison can support a changed-hectares claim.
Expected intervention: identify the phenology mismatch; withhold the hectare claim; request matched-season imagery or a season-aware time series; require registration, mutual masks, sensitivity analysis, and error-adjusted area uncertainty.
Make spatial SQL measurable and index-safe
Design production-safe PostGIS SQL to return flooded area in hectares per parcel for two large EPSG:4326 polygon tables. Include invalid-geometry handling and an EXPLAIN verification plan.
Expected intervention: select a suitable projected measurement CRS, avoid transforming indexed columns inside predicates, repair derived working geometry, deduplicate overlaps, and verify the query plan.
Use this AHP pairwise matrix to build a solar-suitability map. Its consistency ratio is 0.19.
Expected intervention: reject the weights before mapping, identify discordant judgments, require revision, and preserve sensitivity analysis as a deliverable.
- Trigger narrowly.Specialist descriptions identify the domain decision that owns the task; negative and collision cases test over-triggering.
- Load progressively.The agent reads only the relevantSKILL.md, then loads references or scripts when required.
- Enforce invariants.CRS, validity, leakage, units, uncertainty, provenance, verification, and failure behavior travel across the workflow.
- Route across specialists.The orchestrator coordinates genuine multi-stage work without replacing specialist judgment.
- Fail loudly.Missing inputs produce a narrower claim, provisional plan, clarification, or refusal—not invented evidence.
- Fail-loud spatial computing.Successful execution is not proof of a valid geographic result.
- Methodological honesty.Uncertainty, multiple-testing correction, sensitivity, and error-adjusted estimates are first-class outputs.
- Anti-leakage by default.Spatial and grouped validation replace random splits whenever geographic generalization is claimed.
- Tool-pragmatic, vendor-neutral guidance.Open Python, PostGIS/DuckDB, Earth Engine, ArcGIS, and other backends are selected by evidence and scale.
- Progressive disclosure.Long references and scripts cost no context until the task actually needs them.
- Measured, not assumed.Source cases, run evidence, suite hashes, errors, costs, limitations, and superseded results are kept distinguishable.
geoai-skills/ ├── skills/<skill-name>/ │ ├── SKILL.md │ ├── agents/openai.yaml │ ├── scripts/ │ └── references/ ├── evals/cases/<skill>/ # native development-only cases ├── evals/external/ # separately reported transfer suites ├── tools/ # validators, adapters, builders, and gates ├── benchmarks/ # immutable published evidence packages ├── .codex-plugin/ # OpenAI skills-only plugin manifest ├── .claude-plugin/ # Claude marketplace and plugin manifests ├── BENCHMARK.md # published result card and limitations ├── EVALUATION.md # provider-neutral evaluation protocol ├── CASE_STUDIES.md # evidence policy and accepted cases └── RELEASING.md # release and rollback runbook
Contributions are welcome. Start withCONTRIBUTING.md. A skill change should:
- pass the structural and link validators;
- add or update evaluation cases;
- avoid duplicating canonical cross-cutting rules;
- define verification and failure behavior; and
- keep runtime packages free of benchmark and private development material.
Real-world catches belong inCASE_STUDIES.mdonly when they include a dated, privacy-safe reproducer, comparison evidence, verification method, and limitations.
The 1Password MCP server creates a bridge that allows MCP clients such as Codex and Kiro to manage your 1Password Environments with secure authorization prompts.
This is the 1st, easiest, and cheapest PPT, slides, presentation AI generation MCP Server in the world.
Persistent memory for any AI assistant. Zero token cost until recall. Stores memories in local SQLite, ranks by 6-factor scoring, returns results 79% smaller than JSON. Works with Claude, ChatGPT, Grok, Cursor, Windsurf, and any MCP client.
A MCP server that enables AI assistants to interact with Anki, the spaced repetition flashcard application.
Enables LLM clients to interact with macOS applications through AppleScript. Built using the @beyondbetter/bb-mcp-server library, this server provides safe, controlled execution of predefined scripts with optional support for arbitrary script execution.
An MCP server for WordPress plugin audits
Turn your AI assistant into a digital marketing hub that creates, organizes, and analyzes links and QR Codes on demand.
Connect AI clients to Cal.com scheduling through the Model Context Protocol using the hosted server at mcp.cal.com or a local instance.
Sync Calendars, Scheduling Links, AI Executive Scheduling Assistant, Unified Calendar
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
