Know your data sources.

Clearfork Intelligence — curated Texas & New Mexico oil & gas well data, TX drilling permits, and monthly production history with enrichment (aggregated production stats, SB-1150 "zombie well" exposure, canonical well IDs, radius + batch lookup). A cleaned commercial layer complementing the public rrc/nmocd connectors. AUTH: PAID Bearer token (key format cfork_live_...). Set CLEARFORK_BEARER_TOKEN. Requires an Operator ($249/mo) or Enterprise ($499/mo) plan; the permits methods are Enterprise-only. getServiceInfo works without a key (hits the public /health endpoint).

Cognite Data Fusion (CDF) connector for industrial assets, events, and time series data.

Energy intelligence platform with wells, rigs, and production endpoints.

iWell field-data platform: pumper data, tanks, meters, production, run tickets, well groups, custom fields, validation, notes. Auth: OAuth2 password grant (client_id, client_secret, username, password). Token cached 7 days. Rate: 24 req/min. Date params: `since` = UNIX epoch seconds (30-day lookback). `start`/`end` = YYYY-MM-DD inclusive, 30-day window max (server enforces 30 even though docs say 31). `hourly=true` on tank readings requires start===end. Numeric updated_by/author fields → call list_users_map for {id:name} dict. 74 read+write methods. DELETEs omitted for safety.

Palantir Foundry — enterprise data platform. Ontology, datasets, SQL, orchestration, admin, AIP agents, streams. Domains: Ontology (objects/links/actions/queries/timeseries/attachments), Datasets+Branches+Transactions+Files, SQL (async), Orchestration (builds/schedules), Admin (users/groups), AIP Agents, Streams, Connectivity, Filesystem. RULES: 1. Requires Bearer token (PAT) from {stack}/workspace/settings/tokens. 2. All methods need 'stack' (e.g. "mycompany.palantirfoundry.com"). 3. Ontology operations need 'ontologyRid' — use listOntologies. 4. Use listObjectTypes before querying objects. 5. Search filters: {field, type:"eq", value}. 6. SQL is async: executeSql → poll getSqlQueryStatus → getSqlQueryResults. 7. Builds async: createBuild → poll getBuild.

SafetyCulture (iAuditor) connector for audits/inspections and task workflows.

Samsara fleet/IoT — vehicles, equipment, drivers, GPS, telematics, HOS, safety, geofences, DVIRs. AUTH: Bearer token. Cursor-based pagination via pagination.endCursor → 'after' param. RATE: 150 req/s/token, 200 req/s/org. HTTP 429 with Retry-After. PATTERNS: snapshot (current), feed (cursor polling), history (time-range). STAT TYPES (max 3): gps, engineStates, engineRpm, fuelPercents, obdOdometerMeters, batteryMilliVolts, faultCodes, etc.

Snowflake SQL API for query execution and metadata discovery. RULES: 1. Warehouse REQUIRED: { warehouse: "COMPUTE_WH" } or set SNOWFLAKE_WAREHOUSE. 2. Discovery flow: listDatabases → executeSQL (schemas) → listTables → describeTable → executeSQL (data). 3. Use describeTable to verify column names before querying. 4. Use DB.SCHEMA.TABLE when not setting database/schema context.

Spire Maritime vessel data — vessel registry, port events by vessel or location, port details, and voyage tracking from Spire's AIS satellite network.

WattTime marginal emissions signals and forecasts — real-time MOER (lbs CO₂/MWh), historical data, and forward-looking forecasts by grid region.

Whitson+ reservoir engineering. PVT, DCA, RTA, BHP, history matching, nodal analysis, production, well monitoring. 245 methods covering the full Whitson+ API. Auth: Bearer JWT token.