Research2 Buyer Diligence

Buyer-Side Technical Due Diligence in the Voluntary Carbon Market

How major carbon-credit buyers actually evaluate a credit before they wire money — and what that implies for software serving LATAM project developers.

Voluntary-carbon-market (VCM) buyers in 2024–2026 no longer “trust the registry.” After the Guardian/Die Zeit REDD+ scandal of 2023, the rapid rise of independent ratings agencies, and the operationalization of the Integrity Council for the Voluntary Carbon Market (ICVCM) Core Carbon Principles, large corporate purchasers have professionalized their diligence stacks: published criteria PDFs, contracted technical-diligence advisors, dual-rated portfolios, and contractual disqualifiers baked into their RFPs. This section describes the evidence pack every credible LATAM nature-based project must now ship with, the process that turns a project into a purchase, and the operational checklist — ICVCM CCPs, VCMI Claims Code, validation/verification body (VVB) protocols — that gates the whole flow. The implication for project-side software is concrete: the data model, the artifact-vault, and the buyer-facing lineage page are no longer nice-to-haves; they are the unit of diligence.

Part A — Evidence pack: what buyers actually require, by buyer

A.1 Microsoft (advised by Carbon Direct)

Microsoft is the largest single VCM buyer (more than 22 million tonnes contracted as of mid-2025) and runs every short-listed project through Carbon Direct’s technical diligence (Carbon Direct & Microsoft Release 2025 Criteria for High-Quality CDR). The published criteria document — now in its fifth edition (2025 Criteria for High-Quality Carbon Dioxide Removal PDF) — is the de facto industry rubric. The six science-based principles are:

1. Social harms, benefits, and environmental justice
2. Environmental harms and benefits
3. Additionality and baselines
4. Measurement, monitoring, reporting and verification (MMRV)
5. Durability
6. Leakage

(Listed in Microsoft & Carbon Direct, July 2025 release.)

Required artifacts a credit must ship with for Microsoft:

• Project Design Document (PDD) consistent with an ICVCM-eligible methodology
• Validation report by an accredited VVB
• Most recent monitoring/verification report and issuance log on a public registry
• Quantified durability claim with the reversal/buffer mechanism described
• Land-tenure documentation for the project area (cadastral reference, any community-management or “carbon rights” instrument)
• Free, Prior and Informed Consent (FPIC) documentation for any Indigenous Peoples and Local Communities (IP&LC) presence
• A benefit-sharing plan with disclosable financial terms
• Geospatial dataset (project boundary KML/shapefile, leakage belt, plot inventory points)
• Life-cycle assessment / techno-economic analysis where the pathway requires it (BECCS, biochar, mineralization)

Disqualifying conditions (inferable from criteria text and Microsoft’s Procurement Cycle Guidance):

• Methodology not on Microsoft’s eligible list (effectively: not ICVCM CCP-approved or pending approval, and not aligned with the durability principle)
• Vintage too old (Microsoft now favors forward-purchases and recent-vintage issuance)
• Missing or generic FPIC; “implied consent” never qualifies
• No quantified leakage assessment
• No project-specific dynamic baseline for ARR/REDD+ pathways

Diligence vendor: Carbon Direct, exclusively, for technical diligence of nature-based and engineered CDR (e.g., the Re.green Atlantic Forest deal review and the BTG Pactual TIG diligence). Carbon Direct acquired Pachama in November 2025, folding Pachama’s MRV platform into its diligence stack (Carbon Direct acquires Pachama).

A.2 Frontier Climate (Stripe, Alphabet, Shopify, Meta, McKinsey, JPMorgan, Autodesk, H&M)

Frontier funds durable CDR (>1,000-year storage) — relevant to LATAM mostly through biochar, BiCRS, enhanced weathering and ocean-alkalinity pathways. Even if a LATAM nature-based project is not Frontier-eligible, the Frontier rubric is now the canonical model for prepurchase diligence and several pathway-specific public rubrics exist:

• Pathway-agnostic Apply for prepurchase: four lenses — scientific quality, commercial viability, governance, and field advancement
• Marine CDR rubric co-authored with Ramboll
• Enhanced-weathering measurement rubric co-authored with Yale and Georgia Tech (Frontier writing)

Required artifacts (prepurchase track):

• Lab-scale performance + preliminary stability data
• Documented MRV protocol with itemized uncertainty sources and a plan to reduce them
• Environmental and human health & safety risk assessment
• Techno-economic model showing path to <$100/t

Offtake track raises the bar to a pre-FEED-grade techno-economic analysis and a requirement to deliver ≥10,000 tonnes within five years. Every offtake delivery is third-party verified and registry-issued (see Wikipedia for Frontier Climate).

Disqualifiers: durability under ~1,000 years; no externally reviewable MRV plan; no scaling plan beyond the prepurchase volume; storage solution not yet proven at lab scale.

Diligence vendor: Frontier’s in-house team (≈60+ external technical reviewers) plus rotating expert panels — externalized, but Frontier-coordinated. (Stripe Climate Orders documentation describes the same flow.)

A.3 Symbiosis Coalition (Google, Meta, Microsoft, Salesforce — nature-based ARR/agroforestry/mangrove)

Symbiosis is the single most relevant buyer coalition for LATAM nature-based developers: Google, Meta, Microsoft, and Salesforce committed to contract up to 20 Mt of nature-based removals by 2030 (Symbiosis launch announcement). The coalition publishes a 12-criterion Quality Standard (Symbiosis Quality Criteria and the Reforestation & Agroforestry Quality Criteria page) organized around five pillars: conservative accounting, durability, social and community benefits, ecological integrity, and transparency.

The 12 quality criteria (QC1–QC12), each with required evidence:

Disqualifiers per the updated April 2026 Mangrove RFP rules include any project that can’t satisfy ICVCM CCP-aligned methodology, any project missing FPIC, and any project that fails the dynamic performance benchmark.

Diligence vendor: TerraCarbon explicitly thanked for “analysis and support” in drafting; the Technical Advisory Board (Symbiosis TAB) includes Conservation International, The Nature Conservancy, BeZero, Calyx Global, Sylvera, Silvestrum, Yale, UC Berkeley, and TerraCarbon. The 2024 inaugural RFP routed shortlisted projects to TerraCarbon and Xilva for technical evaluation.

A.4 NextGen CDR Facility (South Pole + Mitsubishi)

NextGen aggregates demand from BCG, LGT, Mitsui O.S.K. Lines, Mizuho, Swiss Re, and UBS, focusing on durable removals (BiCRS, DACS, EW, biochar, mineralization) (NextGen overview). NextGen requires:

• ICROA-endorsed standard certification
• Independent third-party MRV (NextGen partnered publicly with Carbonfuture for dMRV, see Carbonfuture case study)
• Operational at scale by 2025; Mt-scale plan by 2030
• Active monitoring of delivery risk by South Pole’s portfolio team

A.5 LEAF Coalition (ART TREES jurisdictional REDD+)

LEAF — the public-private coalition mobilising more than US$1 billion for jurisdictional REDD+ — purchases only TREES Credits issued by ART (ART/LEAF announcement). Required artifacts at the jurisdiction level:

• Crediting-level (FREL) documented per TREES 2.0 (TREES 3.0 in Summer 2025 public comment draft)
• National forest-monitoring system data
• Cancun Safeguards self-assessment using TREES’ 16 themes / 44 indicators
• Benefit-sharing plan or roadmap (LEAF requires it even where ART does not)
• Indigenous Peoples and Local Communities engagement evidence — IP&LC participation is a stated LEAF prerequisite (LEAF IP&LCs page)
• Validation/verification by an ART-approved independent body

Disqualifiers: subnational programs not nested in a national accounting frame; FREL above the TREES allowable benchmark; absence of IP&LC consultation evidence.

A.6 Apple Restore Fund (managed by Climate Asset Management)

Apple’s >US$1B Restore Fund applies a fund-style natural-capital diligence rather than per-credit RFP review. Fund managers (Climate Asset Management for the second tranche; Goldman Sachs/Conservation International for the first) “carefully assess prospective managers and investments to ensure they meet robust environmental, social and governance criteria, and strict standards for quality, scalability, and impact” (Apple Carbon Removal Strategy white paper, Restore Fund expansion announcement). Practically this means audited natural-capital fund accounting on top of standard-issued credits.

A.7 Watershed (corporate buy-side platform, ~500 customers)

Watershed’s 2026 Carbon Removal RFP procures both nature-based and engineered tonnes for its customer network. The platform’s carbon buyer’s guide explicitly states Watershed “vets every decarbonization project for quality, environmental impact, and climate risk” and uses ICVCM recommendations as a signal but not a definitive gate. Watershed runs in-house due diligence; the platform’s marketplace tile lists pricing, vintage, methodology, and bundled diligence notes.

A.8 Salesforce Net Zero Marketplace

The marketplace publicly aggregates third-party ratings from Calyx Global and Sylvera for every listed project (Net Zero Marketplace FAQ PDF, Salesforce launch story). Listing requirements include registered credits, third-party rating coverage, SDG alignment statement, and transparent unit pricing.

A.9 Stripe Climate

Stripe channels purchases through Frontier; its public criteria state buyers fund (1) >1,000-year storage, (2) non-arable-land-bound pathways, (3) potential <$100/t at scale (Stripe Climate Orders docs, Stripe Carbon removal inventory). All deliveries third-party verified.

A.10 Shopify Sustainability Fund

Shopify reviews applications “for climate benefit, scalability, need for voluntary market support, and co-benefits” (Shopify Carbon Removal Application Process). Two portfolios — Frontier (durable) and Evergreen (nature-based + renewables). Portfolio companies are monitored “like a VC fund” — runway, cash burn, financial viability — making project-developer financial transparency an explicit diligence input (Trellis lessons piece).

A.11 JPMorganChase

JPMC published its Carbon Market Principles defining eight evaluation factors: real, measurable, additional, unique/traceable, independently verified, no displacement, durable, equitable. JPMC committed to procure ~800,000 t of durable removals (JPMorganChase 2023 announcement). The October 2025 biodiversity guidance co-published with Carbon Direct adds biodiversity-outcome reporting as an explicit nature-based criterion.

A.12 Klarna

Klarna’s Climate Solutions page documents an “impact-per-dollar” approach via Milkywire’s Climate Transformation Fund rather than an offset register. Differentiated internal carbon fee ($200/t S1+S2, $100/t business travel, $10/t other S3) — diligence is therefore impact-attribution, not credit-quality, but evidence of project additionality and community benefit is still required.

A.13 Buyer × required-artifact matrix (LATAM nature-based focus)

Part B — Diligence process: how they actually do it

B.1 Carbon Direct — the de facto industry template

Carbon Direct’s published redacted diligence write-ups for Microsoft are the closest thing to an open-source industry template. The two most-cited cases — Re.green Atlantic Forest restoration (Brazil) and BTG Pactual Timberland Investment Group (Brazil + Uruguay) — show this section structure:

1. Project overview — scale, geography, methodology, timeline
2. Methodology & accounting analysis — dynamic baseline test, additionality interrogation, leakage handling
3. Carbon quantification & MRV review — allometrics, plot stratification, remote-sensing approach
4. Ecological & ecosystem-impact analysis — species composition (Re.green’s 300-species native list was specifically validated), biodiversity-impact projections
5. Social & community impacts — interviews with local stakeholders, FPIC review, benefit-sharing analysis
6. Operational risk & delivery risk — site visits to plantations and nurseries, tenure verification, financial-capacity check
7. Commercial de-risking analysis — capitalization adequacy, contract structures, recommended risk mitigants

Carbon Direct staffs about 70+ scientists (Carbon Direct Science Team) and reports having conducted 600+ project diligences and 150+ deep technology reviews (Technical Diligence service page). Service offerings are tiered: rapid red-flag analysis → desk-based review → extended iterative review including site visit. Site visits typically pair a forest scientist or carbon scientist with in-country experts and consist of multi-day plantation/nursery walks plus structured interviews with project staff, stakeholders, and local communities.

B.2 TerraCarbon

TerraCarbon is the technical analysis partner credited in the Symbiosis Quality Criteria (Symbiosis updates announcement). They run validation, methodology development, and project diligence services and operate as a Verra-approved consultant. Their internal process is private but their output for Symbiosis is the public benchmark.

B.3 Xilva

Xilva (Switzerland) provides forest-asset due diligence with a heavy field-component bias and is also a Symbiosis-affiliated reviewer. Their process integrates remote-sensing baseline with on-ground forest cruises.

B.4 Sylvera

Sylvera’s published Carbon Credit Framework whitepaper and How we rate carbon credits describe a proprietary, project-type-specific framework with three pillars: Carbon, Additionality, Permanence. Pillars are combined in a matrix so under-performance in one cannot be hidden by over-performance in another. For ARR projects (Sylvera ARR ratings approach), Sylvera uses canopy-height as a proxy for forest growth/loss via proprietary ML models on satellite imagery. Outputs: ratings A–D (with AAA at top edge depending on segment), plus a per-pillar score, plus a quantitative Carbon Score reflecting the share of issued credits the project can demonstrably back. Distribution: integrated into the Sylvera platform via API, and via BlueLayer’s API integration connecting 200+ project developers to Sylvera Connect to Supply.

B.5 BeZero

BeZero’s methodology assessment framework decomposes ratings into six risk factors: Additionality, Over-crediting, Non-permanence, Leakage, Perverse incentives, Policy & political environment. Over-crediting is decomposed across baseline carbon stock, project carbon stock, leakage emissions, and non-permanence deductions (BeZero over-crediting series). BeZero’s published statistic — over 70% of rated forestry projects show high over-crediting risk, driven by the difficulty of estimating counterfactual baselines — is itself a pricing and procurement input for buyers.

B.6 Calyx Global

Calyx’s peer-reviewed framework applies a three-part program-, methodology- and project-level evaluation across six risk factors (additionality, baseline, project emissions, leakage, permanence, overlapping claims) plus a 10-area / 55-sub-area safeguards analysis (Calyx ratings approach explained). Frameworks are externally peer-reviewed by GHG and SDG expert panels.

B.7 Renoster

Renoster’s open-source approach uses implicit deforestation tests via reference areas in the same municipalities (Carbon Market Watch comparison report). Their report “Fixing Verra” proposes structural reforms: random VVB assignment (jury model), shifting auditor payment to a global fund, and detection-bonus compensation.

B.8 MSCI Carbon Markets (ex-Trove)

MSCI (Carbon Project Ratings) covers 4,000+ projects on six axes — additionality, quantification, permanence, co-benefits, reputational risk, delivery risk — on an AAA–CCC scale. Per the 2024 VCM in Review, fewer than 10% of projects rate AAA–A, an indictment of the standing inventory.

B.9 CCQI

The Carbon Credit Quality Initiative (Environmental Defense Fund + Öko-Institut + WWF-US) scores carbon-credit types on seven quality objectives: (1) robust GHG impact, (2) avoiding double counting, (3) addressing non-permanence, (4) avoiding negative environmental impacts, (5) social impacts, (6) host-country policy alignment, (7) facilitating transition to net-zero compatible activity. Each objective is scored 1–5; CCQI explicitly does not aggregate to a single score, leaving weighting to the buyer (CCQI scoring page).

B.10 Cycle time and team composition

Public material is sparse on diligence timelines, but inferable patterns:

• Frontier prepurchase: 3–4 months between pre-application and announcement, ~3 weeks for the full application after invitation, ~3 days for applicants to respond to anonymized reviewer comments (Frontier prepurchase apply page).
• Carbon Direct full diligence (e.g., Re.green): multi-week to multi-month, including a multi-day site visit. Deliverable is a multi-section comprehensive report with operational, scientific, social, and commercial recommendations.
• A typical buyer-side team mixes a carbon scientist / ecologist (allometrics, species composition, baseline science), a geospatial analyst (remote-sensing detection + boundary verification), a methodology / accounting expert (dynamic baseline, leakage), a community/safeguards specialist (FPIC, benefit-sharing review), a lawyer (carbon rights, tenure, contract risk), and a commercial PM (capitalization, delivery risk, offtake structuring). Carbon Direct’s hiring patterns (Forest Carbon Geospatial Scientist) confirm this composition.

Part C — Auditor / VVB process: the verification step

C.1 Accreditation pathways

VVB accreditation is the gating step. The relevant pathways for LATAM nature-based projects:

• Verra-approved VVB list — Verra publishes the maintained list at verra.org/validation-verification. VVBs accredited under ISO 14065 or under the CDM-DOE legacy can apply. SGS was added in May 2025 (SGS Verra approval).
• Gold Standard VVB list — published at Gold Standard VVB directory.
• ART approved validators for TREES.
• Cercarbono approved Conformity Assessment Bodies for Cercarbono and ProClima (Cercarbono Validation and Verification, SustainCERT Cercarbono page).
• Article 6.4 DOE (UNFCCC) and JCM Third-Party Entity (Japan) — relevant where projects are also targeting international transfers under Paris Article 6.

C.2 Auditor × accreditation matrix (LATAM-relevant subset)

(All “Yes” entries cross-validated against Verra’s VVB program rules and approved list and Gold Standard’s directory; LATAM presence is inferred from public job postings, project signed-off, and corporate office maps.)

C.3 Site-visit pattern

The Verra VCS Validation and Verification Manual (Manual v3.2 PDF) prescribes an opening meeting → desk review → site visit → closing meeting cycle. In practice, a typical LATAM REDD+ or ARR validation/verification engagement involves:

• Desk review (5–15 working days): PDD, methodology applicability, baseline calculations, monitoring plan, evidence files, registry status
• Site visit (3–7 days): boundary walk, plot-sampling verification, nursery and intervention-site inspection, stakeholder/community interviews, document inspection (FPIC, benefit-sharing, tenure)
• Sample-based plot re-measurement: the VVB selects a sample of permanent monitoring plots (frequency varies by methodology — typical ~10–15% of plots) for independent re-measurement
• Findings documentation: Corrective Action Requests (CARs), Clarification Requests (CLs), and Forward Action Requests (FARs)
• Closing & Final Validation Report: all CARs and CLs must be closed before issuance (VCS Manual on findings)

C.4 Common findings — typical non-conformities

From parsing public Verra validation reports and Verra’s own non-conformance literature (Corrections to JNR Validation and Verification Process):

• Misapplied non-permanence risk tool (wrong cash-flow breakeven date — see the VCS Validation/Verification Manual case example)
• Inadequate or undocumented FPIC for affected IP&LCs
• Missing or unverifiable land-tenure documentation
• Plot-sampling stratification errors, off-spec allometric equations
• Boundary discrepancies between PDD shapefile and on-the-ground reality
• Leakage belt mis-defined or leakage discount under-applied
• Stakeholder consultation logs incomplete
• Buffer-pool calculation arithmetic errors

C.5 The conflict-of-interest debate

The structural conflict — the project developer pays the auditor — has become the central integrity issue of 2024–2026. Two studies define the current debate:

• Probst et al. (2024) Nature Communications — review of 2,346 projects found over 84% of issued credits did not reflect real emission reductions (cited by Mongabay’s September 2025 reporting).
• University of Pennsylvania Carey Law School / SSRN (Sept 2025) — examined 95 already-flagged Verra projects; two-thirds of accredited auditors failed to identify the flaws. The study explicitly named SGS, DNV, and Bureau Veritas. Cynthia Giles: “The entire system is structured to maximize the number of credits.”
• Carbon Market Watch (July 2025): “Carbon market auditing process inherently flawed.”

Renoster’s reform proposals from Fixing Verra:

1. Random VVB assignment from a vetted pool (jury model)
2. Auditor payment by a blind pool funded by issuance fees, not by developer
3. Performance bonus to auditors who correctly flag over-crediting
4. Mandatory post-issuance re-audit by a different VVB at year 5

Verra’s response acknowledged feedback but rejected random assignment on cost grounds. The implication for buyers: many now layer their own review (ratings agency + Carbon Direct or equivalent) on top of the VVB report.

C.6 Where software fits in the auditor workflow

Auditors specifically need (and currently lack) the following in usable form:

• Chain-of-custody for every measurement — who measured what, when, with what device, with what calibration
• Evidence vault — versioned, hash-stamped storage of FPIC documents, tenure records, plot-inventory raw files, drone or satellite scenes, allometric equation provenance, monitoring-report artefacts, attribution to a specific page of a specific PDF
• Sampling logs — defensible, reproducible plot selection (random seed, stratification spec, GPS-stamped plot visits)
• Finding tracker — CARs/CLs/FARs as first-class objects with response evidence, status, and audit trail
• Read-only auditor view — a project-side OS exposing a VVB-scoped view that doesn’t reveal commercial information
• API into VVB tooling — most VVBs still operate on Word/Excel/email; even a structured PDF export with embedded JSON metadata is a step up

This is the most under-served surface in current carbon software, and it is exactly where a project-side OS can pull weight.

Part D — ICVCM CCP Assessment Framework: operational decomposition

D.1 Two-tier assessment structure

The ICVCM Assessment Framework (Assessment Framework page, CCP Book V1.1 PDF) operates at two levels:

• Programme-level assessment — evaluates carbon-crediting programmes (Verra, Gold Standard, ART, Climate Action Reserve, ACR, Isometric, Cercarbono, etc.) on governance, tracking, transparency, and robust independent third-party validation and verification.
• Methodology / Category assessment — evaluates specific methodologies (or categories of credits) on additionality, permanence, robust quantification, no double counting, and sustainable development benefits and safeguards.

A credit can carry the CCP label only if both its programme and its methodology have passed.

D.2 What programme assessors check

The full Assessment Framework (Section 4 PDF) lists evidence categories per CCP:

• Effective governance — board independence, conflict-of-interest policies, decision-making procedures, complaint and grievance mechanism, evidence of implementation
• Tracking — registry uniqueness, retirement, cancellation, transfers, anti-double-issuance controls, public APIs
• Transparency — public availability of PDD, monitoring reports, validation reports, audit findings, methodology documents
• Robust independent third-party validation and verification — VVB accreditation procedure, conflict-of-interest rules, rotation, sanctions register

D.3 What methodology assessors check

For nature-based pathways, the Methodology Assessment focuses on:

• Additionality — regulatory-surplus test, financial barrier test, common-practice test, dynamic-baseline plausibility
• Robust quantification — measurement protocols, allometric equations, uncertainty propagation, leakage accounting, baseline construction
• Permanence — buffer pool sizing, monitoring frequency, reversal-risk treatment (high-risk reversal categories must commit to ≥40-year monitoring and ≥20% buffer-pool contribution, per ICVCM permanence guidance)
• No double counting — corresponding-adjustment treatment under Article 6, registry uniqueness
• Sustainable development benefits and safeguards — required in the methodology text, with measurable indicators

D.4 Approvals as of November 2025

From the CCP Impact Report 2025:

• 7 programmes approved (covering ~98% of market volume)
• 38 methodologies approved, 22 explicitly rejected
• 51M+ CCP-labelled credits outstanding by October 2025 (~4% of 2024 market volume)
• CCP-labelled credits trade at ~25% premium
• No renewable-energy methodology has been approved for CCP labelling
• Recent approvals include three REDD+ methodologies, IFM on Non-Federal US Forestlands v2.0, Isometric ISM Reforestation Protocol v1.1, Verra VM0047 (ARR using dynamic baselines) v1 and v1.1, and rice-cultivation methane methodologies (ICVCM August 2025 reforestation-IFM-rice methodology decisions).

D.5 Implication for a CCP-aspirant LATAM nature-based project

The methodology must produce, at issuance, a structured artifact bundle that an assessor could machine-read:

• Baseline construction transcript (data sources, statistical method, dynamic-benchmark control selection, version)
• Additionality test outputs with supporting documents
• Plot-by-plot inventory with measurement metadata, allometric equation, uncertainty
• Buffer-pool contribution log
• Monitoring schedule with future-dated commitments
• Safeguards self-assessment with indicator data
• Registry-unique serial numbers with no overlap to compliance schemes (the no-double-counting test)

D.6 VCMI cross-references

VCMI’s Claims Code of Practice sits on top of ICVCM. Its scoring inputs are the Foundational Criteria (an absolute prerequisite — disclosed inventory, science-based target, transparent reporting) and the Carbon Credit Use rules (Silver: 10–<50% of remaining S1+S2+S3 emissions covered by high-integrity credits; Gold: 50–<100%; Platinum: ≥100%). The Monitoring, Reporting and Assurance (MRA) Framework requires limited-assurance third-party assurance on the claim package. So the diligence chain is now: buyer corporate claim (VCMI) ← credit (CCP-labelled, ICVCM) ← methodology + programme (CCP-approved) ← VVB report ← project artifacts. Each link is auditable; each link demands data the project-side OS must produce.

Part E — What this means for software: synthesis

E.1 The “evidence pack” data model

A credible buyer-facing artifact pack needs, minimally, the following object types:

Hash-stamping (SHA-256 or content-addressed storage) addresses the chain-of-custody gap that the September 2025 Penn Carey study highlighted as systemic.

E.2 Auto-disqualifying gates a project-side OS should enforce pre-issuance

Before an issuance request is sent to the registry, the OS should refuse to compile the package if any of these conditions are unmet:

1. Methodology version not on the ICVCM CCP-eligible list (or pending eligibility, with a documented reason)
2. PDD geometry doesn’t match boundary shapefile within tolerance
3. FPIC documentation missing for any community within or adjacent to the boundary
4. Tenure document missing for any parcel
5. Benefit-sharing plan missing or financial terms blank
6. Plot inventory undersampled vs. methodology stratification rule
7. Allometric equation outside cited range of applicability
8. Baseline construction lacks dynamic-benchmark control selection record (for ARR/REDD+)
9. Buffer-pool contribution below threshold
10. Open finding (CAR/CL) not closed
11. Monitoring period gap >12 months without justification
12. Vintage older than the buyer-coalition-specified cut-off (Symbiosis: vintage from project start, no aged credits)
13. No third-party rating coverage if the target buyer (Salesforce NZM, Watershed) requires one

These are deterministic pre-issuance gates — exactly the kind of thing software does well.

E.3 The buyer-facing lineage page

For each credit (or batch), the developer-side OS should expose a public lineage page that includes (at minimum):

• Registry serial range
• Methodology + version
• Vintage
• Issuance date
• VVB name and report link with content hash
• Programme + ICVCM CCP status
• Buffer-pool contribution
• Boundary map (interactive)
• Plot inventory summary statistics
• Last-monitoring-event date
• Co-benefit indicator data (CCB / SD VISta, GBF targets if claimed)
• FPIC summary (community names, consent dates, document hashes — not the documents themselves to protect privacy)
• Third-party ratings (Sylvera, Calyx, BeZero, MSCI) with last-refreshed timestamp
• Findings register (open + closed CARs/CLs)
• Audit-trail of every artifact change with hash + signer

This is what corporate diligence teams currently assemble manually, in PowerPoint, from disparate sources. The first project-side OS to ship it natively wins major time-to-purchase advantage.

E.4 Auditor-collaboration affordances

Specifically:

• Read-only role with VVB scope (only methodology-relevant data, no commercial pricing)
• Sampling tool: the VVB picks plots; the OS records the seed and reproduces the selection
• Field-visit companion (mobile): GPS-stamped photo capture, voice memo, plot re-measurement form
• Finding tracker: VVB raises a CAR/CL/FAR with attached evidence; project responds with evidence; status flips with audit trail
• Evidence-of-correction: when a non-conformity is closed, the OS records the closing artifact with hash and timestamp

E.5 Integration touchpoints — which vendors and standards have APIs

• Sylvera: REST API for ratings + the Connect to Supply data feed; integrated via BlueLayer’s API
• BlueLayer: REST + webhooks to registries, CRMs, ERPs, and Sylvera
• Carbon Direct: post-Pachama acquisition, exposes MRV platform integrations to its diligence workflow
• Verra registry: public read API for issuance/retirement; document download via web (no rich API)
• Gold Standard registry: public read API
• ART registry: public read; manual doc download
• Salesforce Net Zero Marketplace: ingests rating data from Sylvera and Calyx; not yet a public listing API
• Watershed: ingests credit metadata for corporate-buyer accounting; partnership-only ingestion path
• Persefoni: ingests credit metadata for inventory accounting; partnership API
• VCMI Claims Reporting Platform: corporate users submit claim packages; no public ingestion API for project-side software (developers must produce a package the buyer’s assurance provider can verify)

The strategic implication: an LATAM project-side OS should publish a structured machine-readable lineage page (JSON-LD + signed artifacts) that ratings agencies, buyer ESG platforms, and assurance providers can harvest. The protocol is not yet standardized — first-mover advantage is real.

E.6 Net synthesis for a LATAM project-side OS

The de facto buyer diligence stack of 2024–2026 has converged on roughly the same set of evidence requirements: ICVCM-eligible methodology, dynamic baseline, FPIC, tenure, benefit-sharing, MRV with chain-of-custody, durable buffer-pool, third-party ratings, VVB findings closed. Nature-based LATAM projects in particular are scrutinized at a higher bar than they used to be — the 2023 REDD+ scandal and the September 2025 Mongabay-cited auditor study put nature-based credibility under permanent pressure. A project-side operating system that natively produces the artifact pack (with hashes and audit trail), enforces auto-disqualifying gates pre-issuance, exposes a public lineage page, and provides a read-only auditor view will compress diligence cycle time from months to weeks and materially reduce buyer-side discount-for-uncertainty. That is the wedge.

Key citations

Buyer-published criteria

• Microsoft & Carbon Direct, 2025 Criteria for High-Quality Carbon Dioxide Removal — PDF, web edition, release announcement
• Microsoft, Carbon Dioxide Removal Procurement Cycle: FY25 Guidance Document — PDF
• Carbon Direct, Re.green technical diligence write-up — link
• Carbon Direct, BTG Pactual Timberland Investment Group diligence — link
• Carbon Direct, How technical diligence works — link
• Frontier Climate, Apply for prepurchase — link; Apply for offtake — link; Writing (rubrics) — link
• Symbiosis Coalition, Quality Criteria — link; Reforestation & Agroforestry Quality Criteria — link; Mangrove update — link; Technical Advisory Board — link; Coalition launch — Businesswire
• NextGen CDR — website; Carbonfuture case study; South Pole launch announcement
• LEAF Coalition / ART — IP&LCs page — link; TREES 3.0 public draft (July 2025) — PDF; $1B mobilization — ART announcement; Forest Trends LEAF + TREES pamphlet — PDF
• Apple, Carbon Removal Strategy white paper — PDF; Restore Fund expansion — Apple newsroom; CAM expansion — Climate Asset Management
• Watershed, 2026 Carbon Removal RFP — link; Buyer’s guide — link; Marketplace — link
• Salesforce Net Zero Marketplace, FAQ PDF — link; launch story — Salesforce news
• Stripe Climate, How Climate Orders work — docs; Carbon removal inventory — docs
• Shopify, Sustainability Fund application process — link; Trellis lessons piece — link
• JPMorgan Chase, Carbon Market Principles — PDF; 2023 procurement announcement — link; October 2025 biodiversity guidance with Carbon Direct — link
• Klarna, Climate Solutions — link; Climate Transformation Fund (Milkywire) — link

Diligence vendors / ratings agencies

• Sylvera, Carbon Credit Framework whitepaper — link; ARR ratings approach — link; How we rate — link; BlueLayer integration — link
• BeZero, Methodology assessment framework — link; Over-crediting series — link; Carbon Quality Assessors comparison — link
• Calyx Global, Approach (peer-reviewed framework) — link; Ratings explained — link; State of Quality 2024 — PDF
• MSCI Carbon Markets, Carbon Project Ratings — link; Investment Trends 2024 — link; State of Integrity — link; 2024 VCM in Review — link
• Renoster, Fixing Verra — link; Verra response — link
• CCQI — Methodology — link; Scores — link
• Carbon Direct, Technical Diligence service — link; Science Team — link; Pachama acquisition — link

ICVCM, VCMI, programmes, VVBs

• ICVCM, Assessment Framework — link; Core Carbon Principles — link; CCP Book V1.1 — PDF; Section 4 (Assessment Framework detail) — PDF; Assessment Status — link; CCP Impact Report 2025 — PDF; REDD+ approval announcement — link; Reforestation/IFM/rice approvals — link
• VCMI, Claims Code of Practice 2025 Update — PDF; Claims Code page — link; Explanatory notes — PDF
• Verra, Validation and Verification — link; VCS V&V Manual v3.2 — PDF; JNR corrections — link; Brazil REDD+ rejection — Carbon Pulse
• Gold Standard, V&V Bodies directory — link
• Cercarbono, Validation and Verification — link; Protocol V4.4 PDF — link
• SGS Verra approval (May 2025) — link

Auditor over-valuation and reform debate

• Mongabay, ‘Independent’ auditors overvalue credits of carbon projects, study finds (Sept 2025) — link
• Carbon Market Watch, Carbon market auditing process inherently flawed (July 2025) — link
• Carbon Market Watch & Perspectives, Assessing and comparing carbon credit rating agencies — PDF
• Perspectives Climate Group, Analysis of the ICVCM’s CCPs and Assessment Framework (July 2024) — PDF