Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.
Quelldateien prüfen
Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Ein direkter Befehl überspringt den Prüf-Prompt. Prüfen Sie die Quelle, bevor Sie ihn ausführen.
Query and analyze the full surface of US federal contracting + spending + regulation + partner-vetting + market data. 150 tools across 52 data sources (keyless-first — only 4 sources need a free key). Covers SAM.gov opportunities (active + pre-solicitation, RFP/SOW attachment TEXT, contacting officers), USAspending + FPDS awards/recipients/competition/recompetes, GAO protests, entity vetting (OFAC sanctions, SAM exclusions, FAC single-audit, FDIC bank health, EPA ECHO), SEC EDGAR financials, Federal Register / Regulations.gov / eCFR + FAR/DFARS / Congress / GovInfo, pricing & labor & fiscal (GSA CALC, SCA/DBA wage determinations, BLS CPI/OEWS/QCEW, Treasury, BEA regional GDP, Census CBP market sizing, FRED macro, GSA per-diem), health & research funding (NIH/NSF/ClinicalTrials/CMS/NPPES), cyber compliance (NVD CVE + CISA KEV), trade tariffs (HTS), geo/disaster/state-local open data (Census geocode, FEMA, Socrata, CKAN), lobbying influence (Senate LDA), DOL labor-enforcement, data.gov + SBA size standards. Use for any US federal procurement / spending / regulation / vetting / market-sizing / macro question.
when_to_use
federal contracting, SAM.gov search, GovCon opportunities, pre-solicitation shaping, RFP attachments / read the SOW text, USAspending awards, FPDS award actions, contracting officer lookup, NAICS / PSC analysis, agency spending, recompete radar, incumbent pressure, capture brief, bid no-bid, set-aside contracts (SDVOSB / 8(a) / WOSB / HUBZone), FAR/DFARS clause lookup, FAR compliance matrix, Section 889 / CMMC, vet a firm, OFAC sanctions, debarment / exclusions, single-audit findings, bank health (FDIC), EPA compliance, SEC EDGAR financials, teaming partners, GAO bid protests, wage determination / SCA / DBA / CALC labor rates, BLS CPI escalation / OEWS wages / QCEW market size, Treasury fiscal data, BEA regional GDP, Census business patterns / market sizing, FRED macro (GDP/CPI/rates), GSA per-diem travel cost, Federal Register / Regulations.gov rules, eCFR, Congress bills, GovInfo, grants.gov / federal grants, CFDA, NIH / NSF / clinical trials / CMS Open Payments / NPPES, CVE / CISA KEV cyber, HTS tariff, FEMA disasters, Socrata / CKAN open data, Senate LDA lobbying, DOL wage-hour / OSHA enforcement, data.gov datasets, SBA size standard, which API keys are set
disable-model-invocation
false
user-invocable
true
SAM.gov federal-data skill — 150 tools across 52 sources
This skill teaches Claude how to use the sam-gov MCP server (150 tools wrapped from @cliwant/mcp-sam-gov) to answer the full surface of US federal contracting / spending / regulation / vetting / market questions end-to-end — from discovering (and shaping) an opportunity, to reading the actual RFP/SOW documents, running a FAR/DFARS compliance matrix, vetting a firm (sanctions, exclusions, audits, bank health, EPA), sizing a market (establishments + wages + regional GDP), pricing the bid against statutory wage floors, and pulling macro context for escalation.
Setup: the sam-gov MCP server must be reachable. /plugin install cliwant/mcp-sam-gov registers it automatically; otherwise see the repo README. Host prefixing: Claude Code exposes tools as mcp__sam-gov__<tool>; bare MCP hosts use just <tool>. Use whichever your host gives you.
Tool inventory — 150 tools, keyless-first
Keyless-first: 48 of 52 sources need no key. Only 4 sources require a free key (marked 🔑 below): Census CBP (census_business_patterns), FRED (fred_search_series, fred_series_observations), BEA (bea_regional_data), and DOL's data endpoint (dol_get_dataset; DOL's catalog is keyless). Those 4 tools throw an honest config error without a key. A handful of others take an optional free key for higher limits (see the Keys section). The server is the only source of truth — never invent notice IDs, officer names, award amounts, NAICS codes, or regulation citations.
sam_search_shaping — pre-solicitation radar: Sources Sought / Presol / Special Notices before the RFP exists (keyless rows null naics/setAside/place — get those via sam_get_opportunity).
sam_get_opportunity — full detail for ONE notice by 32-char hex noticeId (POCs, deadline, attachments[], body). Call BEFORE bid/no-bid or compliance work.
sam_fetch_description — full SOW/RFP body as plain text (when detail gave a description URL not inline body).
sam_fetch_attachment_text — READ the attachment (RFP/SOW/Q&A/wage PDF+DOCX+HTML) by its attachments[].url. A scanned/image-only or unreadable doc returns text:null + disclosed reason (honest, not a bug).
sam_attachment_url — build the public download URL for an attachment resourceId (303 → signed S3).
sam_lookup_organization — federal-organization id → canonical fullParentPathName.
sam_lookup_notice_fields — batch-fill nulled naics/setAside/PoP/deadline for 1–100 noticeIds from the opt-in GSA daily CSV. OFF by default (returns enabled:false until SAM_GOV_ENABLE_CSV=1); snapshot can lag ~24h.
grants_search — Grants.gov federal grant opportunities (financial assistance, distinct from contracts).
grants_get_opportunity — full grant detail (description, dates, award ceiling, applicant types, CFDA).
usas_search_awards_by_recipient — a recipient's wins in an agency × NAICS slice.
usas_search_subawards — subcontracts on prime awards (supply chain / teaming).
usas_search_recompetes — recompete radar (preferred): PoP ending in a window (default −90d..+18mo), soonest-first, no silent drops. Public signals only.
usas_search_expiring_contracts — DEPRECATED alias of usas_search_recompetes (legacy shape).
usas_analyze_incumbent — per-award incumbent + PUBLIC pressureHints (single_offer / ceiling_nearly_exhausted / hard_stop_no_options) — HINTS, never a composite score.
usas_search_teaming_partners — small-business teaming discovery by socioeconomic cert × NAICS × agency, exclusion-screened (cert is award-derived, not the SBA cert of record).
usas_spending_over_time — contract-spending time series (fiscal_year / quarter / month).
usas_search_agency_spending — top buying agencies for a NAICS / set-aside.
sam_check_exclusions — SAM debarment/exclusion screen by name and/or UEI/CAGE. Empty ≠ "responsible" — only "no ACTIVE match"; a name match isn't identity-proof (verify UEI/CAGE).
sam_integrity_lookup — one-call integrity screen; integrityFlag = excluded or review_fapiis — NEVER "clear" keylessly (FAPIIS has no keyless API); check fapiisUrl.
fac_search_audits — Single Audit summaries from the Federal Audit Clearinghouse.
fac_get_findings — drill into audit-RISK findings for an entity.
fdic_search_institutions — search the FDIC-insured-institution directory.
fdic_institution_financials — quarterly financial time-series for one institution (by cert #).
fdic_risk_ratios — counterparty risk ratios for one institution.
fdic_institution_history — structural-change event log (mergers, charter changes).
fdic_branch_deposits — branch-deposit footprint.
fdic_bank_failures — historical bank failures & assistance transactions.
fdic_industry_summary — industry & state banking-sector annual aggregates.
echo_search_facilities — search EPA-regulated facilities by state with compliance/enforcement screening.
echo_facility_report — EPA ECHO Detailed Facility Report for one facility (by FRS RegistryID).
Financial disclosure — SEC EDGAR (8)
edgar_lookup_cik — company ticker/name → 10-digit SEC CIK.
edgar_company_filings — a company's SEC filings.
edgar_company_facts — curated XBRL financial facts for a filer.
edgar_company_concept — one filer × one XBRL concept × full reported time-series.
edgar_xbrl_frames — cross-filer XBRL cross-section (one concept across all filers for a period).
edgar_full_text_search — full-text search across EDGAR filings, 2001–present.
edgar_filing_index — bulk cross-filer filing index for a quarter.
regulations_search_comments — search public comments on rulemakings.
regulations_get_docket — one Regulations.gov docket by id.
ecfr_search — full-text search across all of CFR (titleNumber=48 for FAR); returns ranked SNIPPETS. For the FULL text of a hit: far_clause_lookup (FAR/DFARS clause) or ecfr_get_section (any other title); or open its ecfrUrl.
ecfr_get_section — the FULL in-force text of one CFR section by citation (e.g. title 2 §200.1 grants, title 29 labor). For FAR/DFARS (title 48) prefer far_clause_lookup.
ecfr_list_titles — all 50 CFR titles + last-amended dates.
far_clause_lookup — authoritative FAR/DFARS clause text + its prescription ("As prescribed in…") for an exact clause number. Carries farOverhaulRisk currency caveat.
far_search — FAR/DFARS-scoped topic search (excludes GSAM, collapses to current in-force version). Feeds far_clause_lookup.
far_compliance_matrix — cited-clause list (1–25) → proposal-ready Section L/M matrix with pass/fail gate flags (Section 889 / limitations on subcontracting / CMMC). A 404 → unresolved, an eCFR-down clause → separate errored bucket; never drops a clause.
gsa_benchmark_labor_rates — GSA CALC awarded-ceiling market band (a min/median/max DISTRIBUTION, fully burdened — do not re-add wrap; not a single price).
sam_search_wage_determinations — find SCA / Davis-Bacon wage determination(s) for a locality (filter coverage sca/dba, state, county; query matches WD number/title, NOT occupation).
sam_get_wage_rates — prevailing-wage + fringe/H&W rate table parsed from a WD + EO minimum-wage floor (check parseConfidence).
bls_timeseries — BLS time series (CPI-U / ECI escalation, PPI, employment) — the escalation layer.
bls_oews_wages — benchmark occupational wages & employment (OEWS) by area × occupation.
treasury_debt_to_penny — daily total US public debt outstanding.
treasury_avg_interest_rates — average interest rate Treasury pays by security type.
treasury_monthly_statement — Monthly Treasury Statement: receipts, outlays, deficit/surplus by month.
treasury_query_dataset — escape-hatch query over 5 confirmed Treasury Fiscal Data datasets.
🔑 bea_regional_data — regional (county/state/MSA) GDP-by-industry + personal income (BEA Regional; requires free BEA_API_KEY — throws without it; a bad key returns HTTP 200 with an Error object → surfaced as invalid_input; suppression codes → null).
🔑 census_business_patterns — establishments / employment / annual payroll by NAICS × geography (Census CBP; requires free CENSUS_API_KEY; suppression sentinels → null, never 0; geoId/naicsCode are strings).
🔑 fred_search_series — search the FRED macro series catalog (GDP/CPI/rates/unemployment/PPI; requires free FRED_API_KEY). Feed id into observations.
🔑 fred_series_observations — time-series observations for a FRED series (requires free FRED_API_KEY; missing value '.' → null, never 0).
gsa_perdiem_rates — federal travel per-diem: monthly lodging (varies seasonally) + M&IE meal caps by city+state OR zip (keyless via DEMO_KEY; ~10 req/hr shared — set DATA_GOV_API_KEY for more).
Health & research funding — NIH · NSF · ClinicalTrials · CMS · NPPES (9)
cve_lookup — NIST NVD CVE records by cveId or keyword/CPE/severity/date, each row JOINED with CISA KEV status (severity + mandated remediation date in one row). not-in-KEV ≠ safe. Optional NVD_API_KEY raises the rate.
cisa_kev_lookup — filter the CISA KEV catalog standalone (binding BOD 22-01 / BOD 26-04 due-dates). Works when NVD is rate-limited. not-in-KEV ≠ safe caveat on every response.
Trade & tariffs — USITC (1)
hts_lookup — US import-tariff classification + duty rates from the USITC Harmonized Tariff Schedule.
Geo, disaster & state/local open data — Census · FEMA · Socrata · CKAN (8)
census_geocode_address — one-line US address → matched address + Census geographies (tract, CD, place).
census_geographies_by_coordinates — longitude/latitude point → Census geographies.
opengov_list_governments — directory of 525+ US state/local governments on OpenGov Procurement (filter by state/query); feed a result's code to the next tool.
opengov_search_solicitations — a government's LIVE public solicitations (title, status [open = accepting], deadline, portal link). Consumes only the anonymous endpoints the public portal itself calls (keyless).
bonfire_list_organizations — curated live-verified 186-org US seed directory of governments on Bonfire (Euna) (filter state/query); feed a result's org to the next tool.
bonfire_search_opportunities — an org's currently-open opportunities via keyless RSS (referenceNumber, name, closeDate, link). The RSS is the complete open set (totalAvailable = the exact open count).
arcgis_hub_discover_datasets — discover ArcGIS Hub datasets by keyword (GIS / infrastructure / permits / zoning / procurement — the SLED layer Socrata & CKAN don't cover). Global open platform → a DISCOVERY aid (publisher surfaced for vetting), not a curated official-source allowlist.
arcgis_feature_query — query rows from a curated allowlist of 23 US-gov ArcGIS REST feature layers: DC OCP PASS live solicitations (~25k); local-gov checkbooks/contracts (Las Vegas, Baltimore, Naperville, Worcester, Topeka, Miami-Dade…); state DOT bid/award registers (TX / AK / IA / OK). Honest match-count total; epoch-ms date disclosure.
Dataset discovery — data.gov (1)
datagov_search_datasets — search the data.gov v4 catalog for federal open datasets across all publishing agencies.
Small business — SBA (1)
sba_size_standard — SBA size standard for a 6-digit NAICS (set-aside eligibility gate). Returns standardType (receipts/employees/assets), normalized threshold, unit, footnote; value is asOf retrieval — re-verify for high-stakes eligibility.
Labor compliance — US DOL (2)
dol_list_datasets — browse the DOL enforcement/compliance dataset catalog (WHD wage-hour, OSHA, OFCCP, ILAB, MSHA…). Keyless. agency/query filtering is client-side. Feed a row's apiUrl + agencyAbbr into the next tool.
🔑 dol_get_dataset — fetch records from ONE DOL dataset (requires free DOL_API_KEY — data endpoint has no keyless tier, throws without it). Records surfaced VERBATIM (envelope key-gated/unverified); totalAvailable null unless the response carries a real count.
Lobbying & influence — US Senate LDA (1)
lda_search_filings — Senate lobbying filings: who is paid HOW MUCH to lobby WHICH agency on WHICH issue (keyless; optional LDA_API_KEY raises the rate). totalAvailable is the API's real ~1.95M-filing match count, not page rows; income/expenses null-never-0.
Server utilities — key discovery (1)
api_key_status — list every key the server can use, required vs optional, signup URL, what it unlocks, and whether each is currently set (a boolean — the value is NEVER shown). The live source of truth for key config.
nws_active_alerts — live National Weather Service alerts (disaster-readiness, pairs with FEMA).
cbp_border_wait_times — live CBP border-crossing wait times.
search_gov_domains — the authoritative CISA get.gov registry (resolve which org owns a .gov; map SLED entities).
SLED local procurement (the state/local/education bid layer) — three complementary keyless surfaces: (1) socrata_query / socrata_discover_datasets now reach 53 curated hosts — state + major-city (NYC City Record, Chicago/SF/LA) + a deep county/city sweep (KC MO, Pittsburgh, Atlanta, Dallas, Baton Rouge, Fulton/Howard/Ramsey/Macoupin/Prince George's counties, MA-Comptroller CTHRU, USAC E-Rate live bids…) — contracts / vendors / checkbook / live solicitations; (2) opengov_* (525+ govs' live solicitations) and bonfire_* (per-org open-opportunity RSS) for the bid-notice feeds Socrata doesn't carry; (3) arcgis_feature_query for gov ArcGIS-REST layers (DC live solicitations, state DOT bid/award registers). For a "who's bidding / what's open in " question, try opengov/bonfire first, then the local Socrata host.
feedback — build a PREFILLED GitHub issue link (bug / feature / wrong_output) for the USER to open & submit (the server never posts). Use when the user reports a problem or wants a missing capability. schema_drift / upstream_unavailable errors also carry a report URL.
Standard workflows
Workflow 1 — Discover + qualify a single opportunity
sam_search_opportunities (NAICS/agency/state) — or sam_search_shaping for pre-RFP Sources Sought/Presol.
Best hit: sam_get_opportunity(noticeId) for POCs + attachments[].
SOW depth: sam_fetch_description for full RFP text.
Read the attachments (Workflow 8).
(Optional) usas_get_agency_profile for issuing-agency context.
Workflow 2 — Competitive landscape ("who wins at agency X")
If agency is an abbreviation: usas_lookup_agency first (canonical name + toptier code).
usas_search_awards with canonical name + NAICS + FY.
Line items: usas_search_individual_awards. Award actions: fpds_search_awards.
Office-level: usas_search_subagency_spending.
(Optional) usas_spending_over_time for the trend.
Workflow 3 — Recompete radar
usas_search_recompetes for agency × NAICS (window −90d..+18mo; filter pscCodes/setAside/minAwardValue). Soonest-first.
Per candidate: usas_get_award_detail (PoP + set-aside + competition) or usas_analyze_incumbent (incumbent + public pressureHints).
Cross-reference sam_search_shaping (same NAICS) to catch pre-RFP shaping while still shapeable.
Workflow 4 — Teaming / supply-chain map
usas_search_individual_awards for prime awards at the target agency.
usas_search_subawards filtered by prime → sub network.
Per sub: usas_get_recipient_profile (alternate names/hierarchy) + sba_size_standard(naics) (small for the set-aside?).
Find NEW partners: usas_search_teaming_partners(cert, naics, agency) — ranked, exclusion-screened (see Workflow 10).
Workflow 5 — Capture brief / agency intelligence
usas_lookup_agency → canonical name + toptier code.
usas_get_agency_profile — mission + scale.
usas_get_agency_budget_function — where the budget goes.
🔑 bea_regional_data(tableName, geoFips, lineCode) — regional GDP-by-industry / personal income (needs BEA_API_KEY).
(Occupational depth) bls_oews_wages(area, occupation) — wage/employment for the labor category.
The establishments/employment/wages + GDP triad; only census_business_patterns and bea_regional_data need a key.
lda_search_filings(agency, issue) — who lobbies the target agency, on what, for how much (keyless).
Narrow by registrantName / clientName to profile a competitor's B2G footprint (income/expenses null-never-0; totalAvailable is the real match count, not page rows).
Workflow 15 — Labor-compliance vetting (NEW)
dol_list_datasets(agency|query) — find the WHD wage-hour / OSHA inspection dataset (keyless catalog); grab a row's apiUrl + agencyAbbr.
🔑 dol_get_dataset(agency, table, filterField, filterValue) — pull the enforcement records for a partner (needs DOL_API_KEY; records surfaced verbatim).
Workflow 16 — Cyber compliance (NEW)
For FedRAMP / CMMC / SBOM component review:
cve_lookup(cveId | keyword) — severity + CISA KEV status joined in one row.
cisa_kev_lookup(cveId | vendorProject) — standalone KEV membership + the binding remediation due-date (works even when NVD is rate-limited). not-in-KEV ≠ safe.
Workflow 17 — Health / research pursuit (NEW)
nih_reporter_search_projects / nsf_search_awards — awarded research grants on a topic (funder/competitive intel).
clinicaltrials_search_studies → clinicaltrials_get_study; clinicaltrials_facet_counts for the distribution.
cms_search_datasets → cms_query_dataset — CMS Open Payments industry→physician payments; nppes_lookup_provider to confirm a provider NPI.
Workflow 18 — Config / which keys are set (NEW)
api_key_status — the live source of truth: every key, required vs optional, signup URL, what it unlocks, and whether it's currently set (boolean; value never shown).
For the 4 required keys not set, point the user to the signup URL (see Keys section). To confirm a key WORKS, call that source's own tool.
Output discipline (anti-hallucination)
Cite tool calls inline: "VA awarded $410M to Booz Allen across 28 contracts in FY26 (usas_search_awards)."
Never invent notice IDs, NAICS codes, recipient names, contract amounts, regulation citations. If a tool returns nothing, say so.
Use autocomplete guards FIRST: NAICS theme → usas_autocomplete_naics; recipient name → usas_autocomplete_recipient; agency abbreviation → usas_lookup_agency.
Quote sparingly from RFP body / regulation text — at most 1 snippet, < 15 words; point to the source URL.
totalAvailable is the real upstream total, not the rows on this page — report it as the match count; page forward when hasMore.
Suppression / missing → null, never 0 — Census/BEA/QCEW/FRED/per-diem/LDA all map a withheld or missing value to null; a genuine 0 is preserved. Never read a null as "$0".
Key-required tools throw without a key — the 4 🔑 tools (Census, FRED×2, BEA, DOL data) return an honest config error, not a fake empty. Check api_key_status.
Notice IDs starting with demo- are fictional demo fixtures; they don't exist on sam.gov.
Common pitfalls
❌ "VA" passed directly as agency to USAspending → nothing. ✅ usas_lookup_agency("VA") first.
❌ Searching opportunities with no ncode/query — useless deluge. ✅ Always narrow.
❌ Quoting an old noticeId from training — they expire. ✅ Always fresh sam_search_opportunities.
❌ sam_fetch_description before sam_get_opportunity — you don't have the noticeId yet.
❌ Confusing usas_search_cfda_spending (grants) with usas_search_psc_spending (contracts).
❌ A NAICS code from training unverified — they get retired. ✅ usas_autocomplete_naics.
❌ Treating sam_fetch_attachment_texttext:null as a bug — it's honest (image-only / unreadable-keyless, reason disclosed).
❌ ecfr_search for an exact FAR clause (e.g. 52.212-4) — mis-ranks GSAM 552.212-4. ✅ far_clause_lookup (exact) / far_search (topic).
❌ Reading ofac_screen_entity / sam_check_exclusions / sam_integrity_lookup empty as "responsible" — empty = "no ACTIVE match"; FAPIIS has no keyless API (never "clear"). ✅ Also check fapiisUrl; verify UEI/CAGE.
❌ Expecting sam_lookup_notice_fields to just work — OFF by default (enabled:false until SAM_GOV_ENABLE_CSV=1). ✅ Few notices → sam_get_opportunity each.
❌ Treating gsa_benchmark_labor_rates as a single price or adding wrap — it's a fully-burdened CEILING distribution.
❌ gao_protest_lookup is complete:false (recent ~25-item feed only), not full history.
❌ Calling a 🔑 tool (Census/FRED/BEA/dol_get_dataset) with no key — it THROWS. ✅ api_key_status first; set the free key.
❌ Reading a Census/BEA/QCEW/FRED null as 0 — it's suppression/missing, never zero.
❌ Reporting a page's row count as the total — totalAvailable is the real upstream match count (e.g. LDA's ~1.95M corpus); page forward on hasMore.
❌ Summing bls_qcew rows across aggregation levels / ownerships — the file mixes them; a do-not-sum note rides every response.
❌ Treating a CVE/component not on CISA KEV as "safe" — KEV is a curated subset of confirmed in-the-wild exploitation; absence ≠ unexploited.
❌ Coercing dol_get_dataset records — the envelope is key-gated/unverified, so rows are surfaced VERBATIM; totalAvailable is null unless the response carries a real count.
Keys & rate limits
Most tools are keyless. Ask the server live with api_key_status — it lists every key, whether required or optional, the free signup URL, what it unlocks, and whether each is currently set (a boolean; the value is never shown). Set keys via the host env block OR a .env file in the server's working directory (auto-loaded; real env wins). Every key below is free.
Required — the source has NO keyless tier, so the tool throws without it:
Env var
Unlocks
Free signup
CENSUS_API_KEY
census_business_patterns
api.census.gov/data/key_signup.html
FRED_API_KEY
fred_search_series, fred_series_observations
fred.stlouisfed.org/docs/api/api_key.html
BEA_API_KEY
bea_regional_data
apps.bea.gov/API/signup/
DOL_API_KEY
dol_get_dataset (the DOL data endpoint; dol_list_datasets catalog is keyless)
dataportal.dol.gov/registration
Optional — only raise a rate limit or unlock one filter (all tools work keyless without them):
Env var
Effect
SAM_GOV_API_KEY
authenticated SAM.gov v2 search + the organizationName filter + full archive (>~12mo)
DATA_GOV_API_KEY
higher limits across api.data.gov sources (Regulations.gov, FAC, NPPES, CMS, data.gov catalog, GSA per-diem) — lifts the shared DEMO_KEY cap
LDA_API_KEY
higher rate on lda_search_filings
BLS_API_KEY
BLS v2 tier (~500/day vs keyless ~25/day) for bls_timeseries (QCEW/CSV path is un-rate-limited regardless)
NVD_API_KEY
higher NVD rate for cve_lookup
SOCRATA_APP_TOKEN
higher Socrata throttling limits
Creating the account at a signup URL is the one manual step; the server automates discovery (api_key_status) and configuration (.env). To confirm a key actually works, call that source's own tool.
Reporting problems & requesting features
This server has a built-in, PULL-only feedback path — use it to help the user improve the tool:
On a schema_drift or upstream_unavailable error, the { ok: false, error } envelope carries a report URL (a prefilled GitHub issue link). If the failure looks real or persistent, offer it: "This looks like the government API changed / is down — you can report it here: <report URL>."
When the user reports a bug, says a result looks wrong, or wants a capability this server lacks, call the feedback tool with kind (bug | feature | wrong_output), optional tool, and a short non-sensitivesummary. It returns a reportUrl for the user to open and submit.
Never post anything yourself — the server only builds the link; the human opens and submits it. The repo is public, so never put secrets, personal data, or sensitive query values in summary (and tell the user to redact them).
