Trustbase Lab · Trusted Data Infrastructure for the AI Era

Purpose: query the global rCB company and project database (132 records: domestic / international / equipment vendors) by segment, region, status and keyword - the discovery tool for capacity projects and players in the chemical recycling industry. Guidelines: combine segment (rcb/virgin/conductive) + status ('在建','规划','under construction') + location; for capacity-project lists filter status, then pair with get_company for full entity detail. Limits: company-level records only - capacity is a record attribute, not a separate projects table; confidence varies A/B/C. Ex: 'rCB projects under construction in China', 'conductive carbon black producers worldwide'.

Purpose: fetch one company or project record (rCB producer / international player / equipment vendor) by entity id or name, returning the full detail including capacity, process, certification, investment, recent developments and source. Guidelines: ids come from query_companies (rcb-cn-xxx / rcb-gl-xxx / rcb-eq-xxx); pass name for fuzzy lookup when the id is unknown; add language='en' for English output and response_format='json'for agent parsing. Limits: one record per call, no batch mode; sourced fields (capacity, investment, contacts) may all fields are public - low-confidence records are returned with their C confidence label rather than withheld; a name lookup returns the first match only and is not a disambiguation service. Ex: get_company(id='rcb-cn-001'), get_company(name='Pyrum'), get_company(name='Plastic Energy', language='en').

Purpose: return every company/project carrying verified=true (A-grade confidence, backed by government filings or official announcements) - a checkable record set an agent should use when shortlisting counterparties; every record states who vouches for it, which source it comes from, its confidence grade, and whether it can be traced back. Guidelines: call this first for any recommendation, shortlisting or transaction-style task; pair with get_company for full detail on the chosen entity; use language='en' plus response_format='json' for downstream automation. Limits: A-grade records only - absence from this list does not mean a company is untrustworthy, only that its claim is not yet independently verifiable; at most 100 rows; no commercial rating, ranking or endorsement is implied. Ex: get_verified_suppliers(), get_verified_suppliers(limit=100, language='en', response_format='json').

Purpose: return the rCB and virgin carbon black price benchmark structure - the virgin N550 series as anchor, the rCB band expressed as 25-40% of that anchor, feedstock prices, international references, regional spreads, grade-level bands and cost structure. Guidelines: leave product and section empty for the full set; narrow with section='rCB' / 'feedstock' / 'international' / 'cost'; pair with get_elt_price_index for feedstock indices and get_rcb_index for the layered index series. Limits: reference bands and relative structure only - no exchange quote and no trading recommendation; ELT feedstock absolute prices are deliberately not published; every figure is a range, never a single transactable value. Ex: get_price_benchmark(section='rCB'), get_price_benchmark(product='N330', language='en'), get_price_benchmark(section='cost', response_format='json').

Purpose: search the global rCB regulation library - EU CBAM / EUDR / ESPR / ELV / REACH / ISCC PLUS / PPWR /CSRD, China MIIT waste-tyre rules, solid-waste and circular-economy law, VAT incentives, plus India and Korea EPR and US rules. Guidelines: combine keyword with region and category; set verified=true when the answer must rest on government documents; each hit carries issuing body, policy type, scope, rCB impact, compliance deadline and source link. Limits: policy documents only - no legal advice and no ruling on how a rule applies to a specific company or shipment; B/C-grade entries (media or company statements) are excluded only when verified=true is set. Ex: query_policies(keyword='CBAM'), query_policies(region='中国', category='税收优惠'), query_policies(keyword='ISCC', language='en').

Purpose: fetch one regulation record in full - issuing body, policy type, scope, impact on the rCB industry, compliance deadline, source link and an AI-readable summary. Guidelines: accepts the id from query_policies(rcb-pol-xxx) or a name/identifier keyword such as 'CBAM' or '2023/956'; run query_policies first when the id is unknown; add response_format='json' for agent pipelines. Limits: one record per call; no legal opinion and no applicability assessment for a given company, product or shipment. Ex: get_policy(id='rcb-pol-001'), get_policy(name='CBAM'), get_policy(name='2023/956', language='en').

Purpose: search the rCB technology and standards library - ASTM D36 committee series (D8178 / D8466 / D8474 /D8632 / D1618 / D1506 / D8491 / D8585), Chinese standards (HG/T 5459 / GB/T 40009 / GB/T 19208 / GB/T 13460 /GB/T 3782 / T/CTRA 01), continuous pyrolysis processes, rCB post-treatment and sustainable carbon black.Guidelines: combine keyword with category (terminology / test / product / process / traceability) and region; use verified=true to keep only entries backed by standard texts. Limits: standard summaries and key technical parameters only - full standard text is never reproduced; C-grade entries are unverified leads; the category taxonomy is ours, not the issuing body's. Ex: query_techs(keyword='TGA'), query_techs(category='测试标准'), query_techs(keyword='ASTM', language='en').

Purpose: fetch one rCB technical standard or process by entity ID or keyword - ASTM D36 series (D8178/D8466/D8474...), Chinese GB/T, HG/T, continuous pyrolysis, sCB - with scope, core indicators, status and source link. Guidelines: pass id (e.g. rcb-tech-001) or name keyword ('D8178','pyrolysis','GB/T 40009'); use when an Agent needs the governing standard behind a parameter or certification claim. Limits: metadata + source link only - not full standard text (copyright); verify currency on the official standards platform before compliance use. Ex: 'ASTM D8178 scope', 'rCB national standard GB'.

Purpose: search the first-hand research archive of the RCB industry WeChat official account - conference coverage, process cases, application boundaries and institution-level market forecasts. Guidelines: keyword matches title/topic/author/scope; filter with category for a single theme; every item is A-grade (first-hand research) and carries publish date, topic, source URL and AI-readable summary. Limits: Chinese-language source material; A-grade here means first-hand, not independently peer-reviewed; coverage is 63 articles from 2026-04 onward, not the account's full history. Ex: query_articles(keyword='Pyrum'), query_articles(category='市场预测·机构观点'), query_articles(keyword='EPDM', language='en').

Purpose: fetch one article record in full - title, publish date, author, account, topic category, source URL, confidence and an AI-readable summary. Guidelines: accepts the id from query_articles (rcb-art-xxx) or a title keyword; run query_articles first when the id is unknown; pair with get_announcements_timeline for a chronological view. Limits: metadata and summary only - the full article body is not served; one record per call. Ex: get_article(id='rcb-art-001'), get_article(name='曼谷大会').

Purpose: search the carbon black grade benchmark library - ASTM D1765 native grades (N110-N990, SAF/ISAF/HAF/FEF/GPF/SRF/MT/FT), conductive specialty benchmarks (acetylene black, Super P, ENSACO 250G, Vulcan XC-72R, Ketjenblack EC-300J/EC-600JD) and rCB types - with 6 standard parameters each (iodine_absorption, dbp_absorption, bet_surface_area, ash_content, heating_loss, sieve_residue_325mesh). Guidelines: combine type='native'/'conductive'/'rcb', category (e.g. 'HAF','FEF'), keyword ('N550','N330','Super P'); use as the anchoring benchmark layer before any rCB vs native comparison; rCB records include vs-N550 gap analysis. Limits: 37 records (29 native + 6 conductive + 2 rCB types) - spec ranges, not lot CoA; ash gap (rCB 8-15% vs N550 <=0.5%) caps rCB at semi-reinforcing uses. Ex: 'N330 standard parameters', 'conductive carbon black for lithium battery', 'rCB vs N550'.

Purpose: fetch one carbon black grade in full - the six ASTM parameters as min/max ranges with their test-method standards, application list, process route, brand owner for conductive grades, and for rCB types a gap analysis against the virgin N550 anchor on iodine / DBP / BET / ash / heating loss / sieve residue. Guidelines: use the semantic id (rcb-grade-n550 / rcb-grade-xc72 / rcb-grade-rcb-general) or the grade name ('N550', 'Super P','rCB-General'); pair with query_grades to discover ids and get_grade_spec for specification detail. ⭐ Scope: this returns the FULL business record (application, process route, brand owner, rCB-vs-N550 gap analysis). For a bare ASTM parameter table only, use get_grade_spec instead. Limits: one grade per call; ranges are compiled from public specifications and are not a supplier's guaranteed certificate of analysis; the gap analysis is relative to N550 only, not to every virgin grade. Ex: get_grade(id='rcb-grade-n550'), get_grade(name='Super P'), get_grade(name='rCB-General', response_format='json').

Purpose: dataset overview - version, record counts per table, entity-ID rules (stable IDs like rcb-cn-001, append-only), confidence grading (A/B/C), verified-supplier counts, the full tool index and site module map. Guidelines: call this FIRST when discovering the server, then route to specific query_* tools; pass language='en' for English output, response_format='json' for Agent processing. Limits: metadata only - no record content; counts refresh with each data release (schema 1.10). Ex: 'what data does this server have', 'list available tools'.

Purpose: substitution economics compute for carbon black - given a native price (vcb_price_input), an rCB price (rcb_price_input) and replacement ratio, returns per-ton saving and optional annualized value. Guidelines: supply your own quotes, or fetch market anchors via get_rcb_index first. Limits: pure arithmetic - no market data access, not a price recommendation. Ex: 'N330 12000 RMB/t vs rCB 8000 RMB/t, ratio 0.3, annual 5000t'.

Purpose: quality benchmark - compare measured rCB sample parameters against a target native carbon black grade (e.g. N550) spec ranges; classify each parameter pass / light deviation / significant deviation. Guidelines: pass target_grade plus rcb_params using EXACT keys - iodine_absorption, dbp_absorption, bet_surface_area, ash_content, heating_loss, sieve_residue_325mesh, ph_value, pour_density (NOT OAN/BET/ash). ideal for grading incoming rCB lots against a native anchor before substitution decisions. Limits: needs measured values as input - no lab data on file; thresholds follow the site's anchored-benchmark methodology, not customer specs. Ex: 'compare rCB sample vs N550'.

Purpose: rCB feedstock (end-of-life tyre) price index in GLOBAL caliber - Tier 1 global categories (scrap tyre collection price / tyre pyrolysis oil / crumb rubber / recovered steel wire / rCB) with region-by-region indices benchmarked to the global weighted average = 1000, plus Tier 2 region-specific categories (EU/US gate fee and TDF; CN rubber fuzz, rubber block, 900-mesh and 24-mesh powder). The first category is the scrap tyre collection price (G-ELT) - the single largest cost item for rCB producers. Guidelines: no arguments returns the global cross-section with region spreads; pass region (EU/US/CN/CA/IN) for that region's index across categories; pass category (G-TPO/G-CRB/G-WIRE/G-RCB) for one category's region comparison. Pair with get_feedstock_overview for arisings and processing paths, and get_rcb_index for layered rCB product prices. Also returns a PUBLIC-SOURCE INDEX SERIES (monthly, base month = 1000) covering the tyre value chain: scrap tyre, 24-mesh crumb rubber, pyrolysis oil, and virgin carbon black grades N220/N330/N550/N660, with category-to-anchor ratios - use it for trend direction and cross-category structure (the virgin carbon-black grades act as the rCB benchmark). This series is index-only and unit-less: it cannot be reverse-engineered into absolute prices. Limits: index values, region spreads and ratios only - absolute prices and source platforms are never disclosed because no official public statistics source exists for ELT feedstock prices worldwide; some regions are excluded from a category when the product caliber is not comparable (e.g. China's bead wire is a finished product vs the recovered mixed wire traded in the EU/US). Ex: 'global tyre pyrolysis oil price index vs China', 'crumb rubber price by region', 'which region has the highest TPO price'.

Purpose: overview of the rCB feedstock module beyond prices - waste tyre arisings and recovery volumes by region (China, EU, US, global) with official sources, the processing-path value ladder (retreading > material recycling > energy recovery > civil engineering > landfill), and the categories covered by the module. Guidelines: pass region (CN/EU/US) to focus the arisings table; use this before quoting feedstock availability or recovery-rate figures, then get_elt_price_index for prices. Limits: arisings and recovery rates are drawn from official and industry-association sources with the statistical body labelled per record - figures from different bodies differ by scope and must not be merged; does not return prices. Ex: 'waste tyre recovery volume China 2025', 'how are end-of-life tyres processed', 'EU tyre recovery rate'.

Purpose: the end-of-life tyre (ELT) collection topic for the rCB supply chain - policy targets, the 2026 standard system, the collection-network build-out, the processing value ladder (retreading > recycled rubber > pyrolysis) and market-size figures, with official sources labelled per record. Guidelines: call with no arguments for the full topic, or pass topic (policy/standard/network/ladder/market) to focus one section; pair with get_elt_price_index (category G-ELT) for collection prices and get_feedstock_overview for arisings. Limits: policy and standard records are sourced from official documents; market-size figures are public-information compilations and labelled as such; does not return prices. Ex: 'China end-of-life tyre recycling policy 2030 target', 'ELT collection network China', 'scrap tyre pyrolysis standards 2026'.

Purpose: native carbon black / rCB price trend direction with official anchors only (no raw quote tables). Guidelines: pass optional segment (domestic/export/import/feedstock/rCB) and period (e.g. 2026); use for directional questions like 'is rCB price rising'. Limits: trend direction and anchors only - no absolute quotes; pair with get_rcb_index for numbers. Ex: 'rCB price trend 2026 CN'.

Purpose: ASTM spec parameter ranges (min/max/unit/test method) for native carbon black grades N110-N990, conductive benchmarks and rCB grades - the standard-parameters reference for carbon black. Guidelines: pass grade (e.g. n330, n550) plus optional type (native/conductive/rcb); use to judge whether an rCB sample falls in spec (iodine_absorption, dbp_absorption, bet_surface_area, ash_content, ph_value). Limits: spec ranges only - not lot-specific CoA data. Ex: 'N330 standard parameters', 'rCB N550 spec vs ASTM'. ⭐ Scope: SPEC-ONLY quick lookup (ASTM min/max/unit/test method). For application, process route, brand owner or rCB-vs-N550 gap analysis, use get_grade instead.

Purpose: four-dimension substitution boundary of rCB vs native carbon black - application limits, regulatory access, certification requirements, material spec constraints. Guidelines: pass dimension in {application, regulatory, certification, material_spec} or omit for all four; call before any substitution feasibility claim (e.g. 'can rCB replace N330 in tire?'). Limits: technical/regulatory possibility only, not commercial viability - pricing belongs to calculate_saving. Ex: 'rCB application limit in seals', 'rCB regulatory access EU'.

Purpose: return a chronological timeline of company announcements across the rCB industry - results, capacity, certification and delivery approvals. Guidelines: filter by region and by date range (YYYY-MM-DD) to isolate a window; use alongside query_companies when tracking capacity build-out, and with query_policies when a regulatory date drives an announcement. Limits: announcement-derived items only - claims are not independently confirmed and full press-release text is not served; coverage depends on what companies have publicly announced, so absence is not evidence of inactivity. Ex: get_announcements_timeline(region='欧洲'), get_announcements_timeline(date_from='2026-01-01', limit=50).

Purpose: search the public knowledge-card collection - atomic sourced facts compiled from statistics, standards, papers and official announcements. Guidelines: keyword matches title and fact fields; combine type and region to narrow; every card returns its fact summary together with the source it was compiled from. Limits: publicly releasable cards only - cards marked excluded (with an excluded reason) are filtered out at export and never returned; each card is a single fact rather than analysis, and carries no recommendation. Ex: query_cards(keyword='ISCC'), query_cards(type='标准解读', region='欧盟'). ⭐ Scope: the LEGACY knowledge-card layer (rCB-centric, small curated set). For multi-domain facts (carbon markets / telecom / health / energy / compute) with confidence grades and as-of dates, prefer query_domain_benchmarks instead - it is the current, larger benchmark layer.

Purpose: rCB price index query - returns the index (ABP), absolute reference price (ARP, USD/ton FOB) and AF/VF attribution decomposition for a grade level, period and region; native-carbon-black-anchored (N550 anchor method). Guidelines: pass grade (P1 refined / P2 standard / P3 industrial, default P2), date (YYYY-MM, blank = latest) and region (CN/EU/NA/SEA, default CN); pair with get_rcb_vcb_spread for discount and get_grade_trend for history. Limits: reference range computed from the anchor method - not a transaction price; VF factor is forward-only (no backfill before 2026-Q3). Ex: 'rCB index P2 CN 2026-09', 'current rCB reference price EU'.

Purpose: anchored spread between rCB and native carbon black - anchoring ratio rho, discount rate and price gap per grade/region/period; the quantitative 'how much cheaper is rCB' indicator. Guidelines: pass date (YYYY-MM), region (CN/EU/NA/SEA), grade (P1/P2/P3); pair with get_grade_trend for direction and calculate_saving for custom quotes. Limits: N550-anchored reference range, not transaction price; single-period snapshot, no forecast. Ex: 'rCB vs N550 discount CN 2026-09'.

Purpose: return the historical series of the layered rCB index for one grade tier - period-by-period assessed benchmark price, period-on-period change, anchor, adjustment factor and verification factor. Guidelines: grade defaults to P2 in the CN region; set start_date and end_date as YYYY-MM to slice a window; pair with get_rcb_index for the current cross-section and get_price_trend for category-level trends. Limits: index points only - dimensionless, non-tradable, and never to be converted into an absolute price; early periods may carry an anchor carried forward from an earlier month, and the output flags when this happens. Ex: get_grade_trend(grade='P2'), get_grade_trend(grade='P1', start_date='2025-12', end_date='2026-09').

Purpose: query global carbon black (HS 280300) bilateral trade flows from UN Comtrade official customs statistics - 1751 records, 8 reporters (CHN/USA/DEU/JPN/KOR/IND/THA/TUR) plus World totals, 2024-2025. Guidelines: pass reporter (ISO3 alpha), partner (ISO3/numeric/name), period (YYYY), flow (X=export/M=import) and limit; unit_price_usd_t and share_pct are derived fields (share excludes World); use for capacity shift, trade-partner concentration and price-gap structure analysis. Limits: customs-reported data only - no commercial quote platforms; unit prices with qty missing are flagged 'missing_qty'; guardrail band 300-8000 USD/t. Ex: 'China carbon black exports 2025', 'top US import partners HS 280300'.

Purpose: aggregate the UN Comtrade carbon black (HS 280300) records for a period - per reporter totals for quantity and value, the weighted unit price, and the top-10 partner countries by quantity. Guidelines: pass period (YYYY) or leave empty to aggregate all periods (2024-2025); weighted price equals total value divided by total quantity; pair with get_trade_partner_ranking to drill into a single reporter. Limits: customs-reported totals only - no commercial quote platform data; World aggregate rows are excluded from partner rankings; unit prices are not comparable across reporters where quantity reporting is incomplete. Ex: get_trade_flow_summary(period='2025'), get_trade_flow_summary(response_format='json').

Purpose: rank partner countries for one reporter, period and flow direction by quantity, returning rank, partner, quantity, value, unit price and share. Guidelines: reporter (ISO3 alpha), period (YYYY) and flow (X =export, M = import) are required; raise top_n above the default 10 (max 50) only when a longer tail is needed.Limits: one reporter-period-flow combination per call; World aggregate rows are excluded; shares are computed within the returned set, not against global trade. Ex: get_trade_partner_ranking(reporter='CHN', period='2025', flow='X'), get_trade_partner_ranking(reporter='USA', period='2024', flow='M', top_n=20).

Purpose: describe the trade flow dataset - source, caliber, unit conversion, derived-field declarations, field table and compliance boundaries - so an agent can cite and interpret it correctly. Guidelines: call this before running any trade query that will be quoted; it states explicitly which fields are customs-reported and which are our own derivations. Limits: UN Comtrade official statistics only - no commercial quote platform data enters this dataset; this tool documents the dataset and returns no trade records itself. Ex: get_trade_dataset_overview(), get_trade_dataset_overview(response_format='json').

Purpose: search the cross-domain benchmark layer - the same verifiable-anchor methodology we apply to rCB, extended to other domains (carbon footprint & markets / telecom & space / health & nutrition / new energy & batteries / compute & energy). Records are policy milestones, official limit values, standards status and dated price snapshots, each with an issuer, an as-of date, a confidence grade and a source_url. Guidelines: filter by domain (carbon / telecom / health / energy / compute), region, category or confidence; every record is meant to be citable - pair with get_domain_benchmark by ID for the full record including the source link. Limits: anchor facts only - no price forecasts (carbon prices are dated snapshots, not live quotes), no medical or dosing advice, no diagnosis; the domain taxonomy is ours, not the issuing body's. Ex: query_domain_benchmarks(domain='carbon'), query_domain_benchmarks(keyword='CBAM'), query_domain_benchmarks(domain='health', confidence='A', language='en'), query_domain_benchmarks(domain='compute').

Purpose: fetch one cross-domain benchmark record by ID or keyword - full summary, issuer, status, as-of date, domain impact, confidence grade and the source URL for citation. Guidelines: pass the ID returned by query_domain_benchmarks (e.g. 'tbm-021') or a distinctive keyword; use this whenever a cross-domain fact is going to be quoted, so the citation carries a verifiable source link. Limits: one record per call; anchor facts only - carbon prices are dated snapshots rather than live quotes, and the health domain provides no diagnosis, treatment or dosing guidance. Ex: get_domain_benchmark(id='tbm-021'), get_domain_benchmark(keyword='CBAM', language='en').

Purpose: the CBAM timeline as a single ordered view - from the founding regulation through the transitional period to the definitive regime and the surrender phase. Guidelines: call this first whenever a question involves CBAM dates, obligations or scope ("when does CBAM start", "when must certificates be surrendered"). Grouped into transitional/legislative, definitive and surrender phases by the as-of date of each source document. Limits: timeline only - dates come from the cited official documents and are not forecasts; the phase grouping is ours, not the EU's official staging. Pair with get_domain_benchmark(id=...) for the source URL of any node. Ex: get_cbam_timeline(), get_cbam_timeline(language='en').

Purpose: one-call overview of the carbon market layer - dated carbon price snapshots (EU ETS auction/settlement, China national ETS), market size statistics and allowance allocation rules, each with its issuer and source. Guidelines: use for "what is the carbon price / how big is the market / how are allowances allocated" style questions; filter by market to narrow to the EU or China. Limits: prices are dated snapshots from the cited sources, NOT live quotes; currencies are not converted, so do not compare across currencies without an FX rate. Ex: get_carbon_market_overview(), get_carbon_market_overview(market='cn'), get_carbon_market_overview(market='eu', language='en'). ⭐ Scope: ONE market at a time - price + market size + allocation rules together. For a side-by-side comparison across markets (who is more expensive, what is the spread), use compare_carbon_prices instead.

Purpose: the carbon accounting standards library - ISO 14064 series, ISO 14067, ISO 14068-1, GHG Protocol and withdrawn standards such as PAS 2050, each with status, as-of date, issuer and confidence grade. Guidelines: use when a question involves which standard governs an organisation-level inventory, a product carbon footprint, or a verification statement; filter status='withdrawn' to check whether a standard still applies. Limits: standards catalogue only - no interpretation of how a standard applies to a specific company. Ex: get_carbon_standards(), get_carbon_standards(status='withdrawn'), get_carbon_standards(keyword='14067').

Purpose: put every carbon price snapshot side by side - market/basis, value, unit, price date and source - and compute the spread within each currency. Guidelines: use for "how do EU and China carbon prices compare" or "what is the spread between auction and secondary prices"; same-currency spreads are returned as computed values. Limits: currencies are NOT converted - the tool deliberately does not produce a cross-currency difference, because that requires an FX rate the dataset does not carry; prices are dated snapshots, not live quotes. Ex: compare_carbon_prices(), compare_carbon_prices(language='en'). ⭐ Scope: SIDE-BY-SIDE comparison only (market/basis/value/unit/date/spread, same currency). For a single-market overview that also covers market size and allowance-allocation rules, use get_carbon_market_overview instead.

Purpose: a transparent CBAM cost estimate - embedded emissions multiplied by a carbon price. Guidelines: supply emission_intensity (tCO2e per tonne of goods) and quantity (tonnes); optionally supply carbon_price, otherwise the latest EU ETS EUA spot snapshot in the dataset is used and cited in the output. Use for order-of-magnitude questions about CBAM exposure. Limits: this is ONLY the emissions-times-price multiplication - it excludes free-allocation deduction, third-country carbon-price deduction and default-value mark-ups, so the result is NOT a declarable amount and is not advice; the emission intensity is your input, not our data. Ex: calculate_cbam_cost(emission_intensity=1.8, quantity=5000), calculate_cbam_cost(emission_intensity=2.1, quantity=1200, carbon_price=85).

Purpose: the EU and China battery regulation timeline in one ordered view - applicability dates, legal bases and amendments, including conditional items. Guidelines: call first for any battery-regulation date question (passport, recycled content, due diligence, carbon footprint). Limits: dates are the applicability dates set in the regulation; items marked conditional apply only after the delegated act enters into force (later of the two triggers). Ex: get_battery_regulation_timeline(), get_battery_regulation_timeline(keyword='护照').

Purpose: recycled-content minimums for cobalt, lead, lithium and nickel, split into the two statutory stages. Guidelines: use for threshold questions; pair with calculate_recycled_content_gap to compare against an actual figure. Limits: thresholds only, by material and stage - no interpretation of how they apply to a specific battery. Ex: get_recycled_content_requirements().

Purpose: battery recycling targets - recycling efficiency, material recovery rates and collection rates, with the year each takes effect. Guidelines: use for target/efficiency questions rather than dates. Limits: published targets only, no industry performance data. Ex: query_battery_recycling_targets().

Purpose: China battery-recycling market statistics - collected volumes, processor counts, NEV output, and modelled future volumes. Guidelines: use when a question needs scale rather than rules; modelled figures are labelled as such in the record names. Limits: China focus; modelled values are not official statistics. Ex: get_battery_market_stats().

Purpose: the 3GPP and ITU-R standards roadmap - 3GPP releases alongside ITU-R IMT-2030 framework milestones and the 6G timetable. Guidelines: use for standard-version and timeline questions. Limits: status and dates only - performance figures live in the standards themselves. Ex: get_standards_roadmap().

Purpose: LEO constellation status - in-orbit counts for Starlink, Qianfan, Guowang, OneWeb and Amazon Leo, plus annual launch totals and planned sizes. Guidelines: use for constellation scale questions. Limits: counts are public catalogue snapshots, not live tracking; the UCS satellite database stopped updating in 2023-05, so check source vintages. Ex: get_constellation_status().

Purpose: spectrum allocations and WRC agenda items - China's frequency plan, trial licences and the WRC-27 items relevant to IMT and direct-to-device. Guidelines: use for spectrum and conference-agenda questions. Limits: agenda items are studies in progress - conclusions only exist in the final conference documents. Ex: query_spectrum_allocations().

Purpose: the 6G core patent landscape by country and by company. Guidelines: use when a question needs the competitive patent picture. Limits: shares follow published tallies and agencies count families or filings differently - compare across sources with care. Ex: get_6g_patent_landscape().

Purpose: telecom and satellite-internet policy timeline, including China's 6G roadmap milestones and satellite-internet policy documents. Guidelines: use for policy-sequence questions. Limits: China focus; roadmap dates are planned, not achieved. Ex: get_telecom_policy_timeline().

Purpose: dietary intake limits and targets - WHO and national figures for fat, sugar, salt, sodium, fibre, fruit and vegetables, by population group. Guidelines: use for limit-value questions. Limits: public classifications and limits for reference only - NOT individual health advice and not for diagnosis or treatment; scopes differ between issuing bodies. Ex: query_dietary_limits().

Purpose: dietary guideline entries issued by WHO, China and the US. Guidelines: use when a question asks what a guideline body recommends. Limits: reference only, not individual advice; the three bodies' scopes differ, so do not merge their numbers. Ex: get_dietary_guidelines(keyword='全谷物').

Purpose: physical activity guidelines by population group (adults, children, older adults, pregnancy). Guidelines: use for activity-volume questions. Limits: reference only, not individual advice; note the children's figure is a daily average and WHO sets no quantified threshold for sedentary behaviour. Ex: query_activity_guidelines().

Purpose: common health myths clarified against official or systematic-review wording, plus supporting evidence notes. Guidelines: use when a claim needs checking. Limits: reference only, not advice; clarifications quote the source wording without extrapolation. Ex: query_health_myths(keyword='蛋黄').

Purpose: health indicator reference values - BMI classes, waist thresholds and body-fat reference points. Guidelines: ALWAYS state which standard is being used. Limits: reference only, not diagnosis; BMI and waist exist in TWO standards (WHO/international vs China) with different cut-offs - never mix them. Ex: get_health_indicators().

Purpose: compute infrastructure scale - rack counts, intelligent-computing capacity, hub and cluster aggregates and the national scheduling platform. Guidelines: use for scale questions. Limits: aggregated to hub, cluster or city level only - no precise facility coordinates are held. Ex: get_compute_infrastructure().

Purpose: PUE benchmarks and green-rating figures for compute facilities. Guidelines: use for efficiency questions; pair with calculate_compute_energy_cost when a load figure is available. Limits: values are published statistics or single-site cases - do not average across differing scopes. Ex: get_pue_benchmarks().

Purpose: compute power consumption statistics, carrying TWO official calibers side by side (National Energy Administration vs CAICT) plus data-centre cost structure. Guidelines: quote the issuing body with the figure. Limits: the two calibers differ by roughly 15% and MUST NOT be mixed - always name the source. Ex: get_power_consumption_stats().

Purpose: green power trading and direct-connection projects - certificate prices graded by production year, traded volumes and flagship projects. Guidelines: use for green-power price and project questions. Limits: project locations are city or park level only; certificate prices are dated snapshots, not live quotes. Ex: query_green_power().

Purpose: compare a given recycled-content figure against the statutory minimum for one material, at both stages. Guidelines: supply material (cobalt / lead / lithium / nickel) and current_pct; thresholds are parsed from the regulation records in the dataset and the source record ID is returned with each row. Limits: one material at a time - this is NOT a compliance finding for a whole battery, and not advice or a declaration basis. Ex: calculate_recycled_content_gap(material='cobalt', current_pct=10).

Purpose: a transparent compute-facility energy and cost estimate - IT load x PUE x hours (x tariff). Guidelines: supply it_power_kw; optionally pue (otherwise a published benchmark from the dataset is used and cited), hours (defaults to a full year) and electricity_price (without it only energy is returned). Limits: excludes non-cooling auxiliary loads, transmission losses, demand and capacity charges, green-power premia and carbon cost; the tariff must come from the caller. Ex: calculate_compute_energy_cost(it_power_kw=1000, electricity_price=0.6).

Purpose: the citable safety-boundary layer for strength training - look up an exercise x condition pair and get a graded verdict (avoid / precaution / safe-with-modification / recommended / ineffective-for-target) together with the verbatim source quote, the section locator and a content hash for each citation. This is the wedge of the fitness domain: it answers 'is this movement contraindicated for this condition' rather than 'what should I train'. Guidelines: pass exercise and/or condition as an id (fit-ex-*, fit-cond-*) or as a keyword in Chinese or English; optionally filter by verdict; every row carries at least one retrievable source with a quote you can check character-for-character. Limits: **no individualised advice** (never 'what should I do'), no diagnosis, no dosing; this is a curated source-backed subset, so the absence of a pair is NOT evidence that the pair is safe; paid standards and copyrighted books are represented by metadata only and their body text is never reproduced. Ex: query_safety_boundaries(exercise='深蹲'), query_safety_boundaries(condition='osteoporosis'), query_safety_boundaries(verdict='avoid', response_format='json'), query_safety_boundaries(exercise='fit-ex-bench-press', condition='fit-cond-uncontrolled-hypertension').

Purpose: fetch one full record from the fitness safety-boundary domain by stable id - an exercise, a condition, or an exercise x condition safety boundary - including every citation as a verbatim quote plus its section locator, retrieval timestamp, source content hash and the record's own content hash, so a downstream agent can cite and verify it without re-fetching. Guidelines: get the id from query_safety_boundaries first (or pass name as a keyword when the match is unambiguous); set include_domain_contract=true to also receive the domain contract (id prefixes, required fields, assertion-kind semantics, source-verification summary, hard boundaries). Every record carries assertion_kind - 'explicit' means the source states the conclusion in plain terms, 'derived' means it was inferred from explicit source facts and the reasoning chain is given. Limits: **no individualised advice** and no diagnosis; paid standards / copyrighted books are metadata only; records without source support are not held here at all. Ex: get_fitness_record(id='fit-ex-back-squat'), get_fitness_record(id='fit-ci-machine-resistance-training-vertebral-fracture-risk'), get_fitness_record(name='卷腹', include_domain_contract=true, language='en').