Installation of high efficiency wood burning cookstoves in Malawi

Integrity

The verification report confirms additionality via an activity test and the VVB has validated additionality (additionality confirmed by the VVB). The project uses a project-specific baseline under VMR0006 and reports a quantified leakage deduction of 0.95%, which are positive features. However, there are many material findings and corrective action requests in the verification record (sampling approach, incomplete baseline assessment, measurement of individual stove consumption, and database gaps) that materially weaken confidence in baseline validity and monitoring; buffer/reserve information is not stated and permanence rationale relies on the category being non-AFOLU rather than robust safeguards.

Transparency

The project names the VVB (Carbon Check (India) Private Ltd.) and provides monitoring and crediting periods, and the monitoring method (annual survey) is stated. Despite that, key parameters are missing from available extracts (buffer pool percentage, baseline reassessment date) and the extracted evidence documents field is 'unknown'. There are substantial contradictions in reported emission reduction figures and other fields which reduce transparency; extraction confidence is medium, which further lowers trust in completeness of the record.

Claim Safety

Corsia eligibility is explicitly false, lowering dual-claim risk on that axis. However, the presence of major discrepancies in ERR (for example contradictions show total_err_claimed 28,795 vs total_err_verified 1,307,556 in later documents) and unresolved CARs about sampling and stove operation create substantial over-crediting and claim safety concerns. Usage monitoring has a verified value (usage rate verified 87.1%) but the assumed usage rate is not stated and the sampling method is disputed, increasing uncertainty.

Documentation

Sixty documents were used and the issuance record is recent (issuance doc dated 2025-11-06), which supports recency. Nonetheless, evidence_docs is listed as unknown, extraction confidence is only medium, and numerous CARs/FARs require revisions to the verification and monitoring reports. The VVB independence appears intact (named VVB), but the VVB is required to revise the verification report and provide detailed assessments — indicating incomplete documentation and outstanding corrective actions.

Overall

I privileged the more recent values in the contradictions list (the 'value_high' entries) because they are tied to later document dates and align with the issuance document (2025-11-06); for example I treated the local-field FNRB method, FPIC conducted = true, annual_survey monitoring, safeguards mentioned = true, and the crediting period 2020-12-01 — 2030-11-30 as the authoritative values. Where emission reduction figures contradict (total ERR claimed 28,795 vs a later total ERR verified 1,307,556), I privileged the later/higher reported verified value as reflecting the most recent assessment but flagged this as a reliability problem. These contradictions and the many outstanding CARs/FARs materially lower scores across dimensions because they indicate unresolved methodological and data reliability issues. Given the medium extraction confidence and unresolved corrective actions, the overall score is modest (3.9) reflecting procedural compliance in some areas but significant residual risk of over-crediting and insufficiently demonstrated monitoring/baseline robustness.