A March 2018 proposal to a United States defense research agency, led by EcoHealth Alliance with the Wuhan Institute of Virology among its partners, proposed to analyse coronavirus spike sequences for furin cleavage sites and, where clear mismatches occurred, to introduce human-specific cleavage sites and evaluate the resulting growth potential; the agency did not fund it.

The cleavage-site passage is verbatim at a confirmed locator and the non-award is confirmed three ways (COV-E027); the margin comment in COV-E028, read whole, evidences intended allocation of assays to Wuhan after an award. Short of established only because the available copy is a third-party archive upload of an OCR-derived scan.

The cleavage-site passage is verbatim at a confirmed locator and the non-award is confirmed three ways (COV-E027). Short of established only because the available copy is a third-party archive upload of an OCR-derived scan.

The cleavage-site passage is verbatim at a confirmed locator and the non-award is confirmed three ways (COV-E027, COV-E028). Short of established only because the available copy is a third-party archive upload of an OCR-derived scan.

The 2018 DEFUSE proposal documents a clear, pre-pandemic intention by EcoHealth Alliance and its Wuhan partners to introduce human-specific cleavage sites into SARS-related coronaviruses. The textual evidence firmly establishes that the exact technique yielding SARS-CoV-2's defining feature was actively contemplated by the relevant actors. The strongest opposing consideration is that DARPA rejected the proposal, and a document outlining intended research is circumstantial evidence of capability, not proof of execution.

The DARPA PREEMPT DEFUSE proposal clearly documents a 2018 intention by a consortium including the Wuhan Institute of Virology to insert human-specific cleavage sites into SARS-related coronaviruses. Multiple independent OCR passes of the archived document confirm the verbatim text at a confirmed locator. The strongest opposing consideration is that the proposal was ultimately unfunded and merely demonstrates contemplated capabilities and intent, rather than execution.

The March 2018 DEFUSE proposal text explicitly details plans to evaluate and introduce human-specific cleavage sites into SARS-related coronaviruses in collaboration with Wuhan partners. This provides verified documentary evidence of intent and capability twenty months before the outbreak. The strongest opposing consideration is that the proposal was ultimately rejected and never funded by DARPA, and contemplating an experiment on paper is not proof of its execution.

The DEFUSE proposal text in COV-E027 unambiguously describes analyzing CoV spikes for furin sites and, where mismatches occur, introducing human-specific cleavage sites, with WIV as a partner, and multiple records confirm the proposal was not funded. The main opposing consideration is provenance: the available copy is a third-party archive scan rather than an official release, though its authenticity is indirectly corroborated by FOIA drafts and congressional usage.

The DEFUSE proposal, as OCR‑verified from the archived scan, contains a passage explicitly proposing to analyze sarbecovirus S genes for furin sites, introduce “appropriate human-specific cleavage sites” where mismatches occur, and test growth, with WIV as a partner and Wuhan listed as a performance site (COV-E027), and a FOIA release confirms drafts and associated comments (COV-E028). The strongest objection concerns provenance — the main copy is a third‑party archive upload and the project was unfunded — but its authenticity is corroborated by US government records and congressional usage, and the claim is limited to what the proposal proposed and to DARPA’s non‑funding, both of which are well supported.

The DEFUSE proposal text, corroborated by FOIA‑released drafts, explicitly describes analyzing SARS‑related spikes for furin sites and inserting human‑specific cleavage sites where mismatches exist, in a multi‑site program including Wuhan, and DARPA’s non‑award is well documented. The residual uncertainty is the lack of an official government‑hosted copy of the final submitted proposal, so this specific PDF relies on chain‑of‑custody triangulation rather than direct provenance.

The S2 proteolytic-cleavage passage is verbatim at a confirmed locator in the March 2018 DEFUSE proposal, recovered by two independent OCR passes that agree on every substantive word, and it proposes introducing human-specific cleavage sites where mismatches occur. Non-award is confirmed three ways, including a negative check of federal award records; draft margin comments, read whole rather than truncated, show a presentational intent to allocate assay work to Wuhan after an award, not a settled concealed plan. The strongest opposing consideration is that the available copy is a third-party archive upload rather than an official release, the proposal was rejected, contemplated backbones are argued to be too distant from SARS-CoV-2, and a document describing intended work is not evidence the work occurred.

The DEFUSE proposal at a confirmed locator states that the team would analyse SARSr-CoV spike sequences for furin cleavage sites and, where clear mismatches occur, introduce appropriate human- specific cleavage sites and evaluate growth potential; two independent OCR passes agree on every substantive word. Non-award is confirmed by the congressional record, contemporaneous reporting, and a negative check of federal award data. The full draft margin comment, often truncated, is a presentational judgment about how to pitch a US-weighted team plus an intention to allocate assays to Wuhan after an award, not concealment of a settled plan. Strongest opposing consideration: the available copy is a third-party archive upload rather than an official release, the text is OCR-derived, the proposal was rejected, and a document describing intended work is not evidence the work occurred — which is why performance is graded separately as C017.

The cleavage-site passage is verbatim in the March 2018 DEFUSE proposal at a confirmed page locator, recovered by two independent OCR passes that agree on every substantive word, and the non-award is confirmed by the congressional record, contemporaneous reporting, and a negative check of federal award data. That is a documented pre-pandemic intention, in a program with a Wuhan partner, to insert the class of feature the pandemic virus carries — more than the zoonotic account predicts, and less than execution. The strongest opposing consideration is that the available copy is a third-party archive upload rather than an official release, the proposal was rejected, and a document describing intended work is not evidence the work occurred.

The cleavage-site passage is verbatim in the proposal at a confirmed locator, read by two independent OCR passes agreeing on every substantive word, and the non-award is confirmed three ways including a negative check of federal award records; authenticity is independently corroborated by draft versions in a federal FOIA production. The claim is carefully bounded to what the document says, and that bound holds. The strongest opposing consideration is provenance: the available copy is a third-party archive upload rather than an authenticated official release, and the text is OCR-derived from an image-only scan — which is why this sits at well supported rather than established, with moderate rather than high confidence. On content: the document is genuinely more than the zoonotic account predicts and considerably less than a mechanism, since a rejected proposal establishes contemplated technique, not performed work.

The cleavage-site passage is verified verbatim at a confirmed locator by two independent OCR passes agreeing on every substantive word, and the non-award is confirmed three separate ways including a negative check of federal award records. The strongest opposing consideration is chain of custody: the available copy is a third-party archive upload, OCR-derived from an image-only scan, rather than an authenticated official release — which is precisely what holds this at well_supported rather than established, though the FOIA-produced drafts corroborate authenticity. Objections that the proposal was unfunded or that its backbones were too distant bear on the claim's implication, which the claim explicitly does not assert; as a claim about what a document says, it is on firm ground.

The cleavage-site passage is verified verbatim at a confirmed locator by two agreeing OCR passes, and the non-award is confirmed three independent ways including a negative check of federal award records. The dossier's own reasoning for withholding "established" is sound: the copy is a third-party archive upload, not an official release, and the text is OCR-derived. The strongest opposing consideration is interpretive rather than documentary — the proposal was rejected, its contemplated backbones are argued too distant to be precursors, and a document is not a deed. Assessed strictly as a claim about what the document says, well supported is right; its value is evidence of intent-on-paper, nothing more.

The cleavage-site passage is verbatim at a confirmed page locator, read by two independent OCR passes agreeing on every substantive word, and the non-award is confirmed three separate ways including a negative check of federal award records. The claim is carefully scoped to what the document says and to the rejection. It falls short of established only because the available copy is a third-party archive upload with no official provenance. The strongest opposing consideration is not authenticity but inference: a rejected funding application describing intended technique is circumstantial evidence about intent and capability, and the contemplated backbones are argued to be too distant from SARS-CoV-2 to serve as precursors.

The cleavage-site passage is verbatim at a confirmed locator, read by two agreeing OCR passes, and the non-award is corroborated three independent ways including a negative federal award check; draft versions appear in an authentic federal FOIA production. The claim asserts only what a document proposed, and it is true at that level. The strongest opposing consideration is chain of custody — the copy read is a third-party archive upload rather than an official release or authenticated production, and the text is OCR-derived — which is why this falls just short of established. The frequently truncated margin comment, read whole, supports a presentational reading about pitching a US-weighted team with intended later allocation of assays to Wuhan, not a settled concealed plan.

The cleavage-site passage is verbatim at a confirmed locator, read by two agreeing OCR passes, and the non-award is confirmed three independent ways including a negative check of federal award records; draft versions appear in an authentic federal FOIA production. The strongest opposing consideration is chain of custody — the available copy is a third-party archive upload rather than an authenticated official release — which is why I stop short of "established." As stated the claim is purely documentary and I grade it as such; its significance is that it makes furin-site insertion a contemplated technique in a program with a Wuhan partner twenty months before the outbreak, which is more than zoonosis predicts and far less than a mechanism.

The cleavage-site passage is verbatim at a confirmed locator and the non-award is confirmed three independent ways. Short of established only because the available copy is a third-party archive upload and the text is OCR-derived.

The cleavage-site passage is verbatim at a confirmed locator and the non-award is confirmed three ways (COV-E027, COV-E028). Short of established only because the available copy is a third-party archive upload of an OCR-derived scan. Every seat treats it as documented intent-on-paper; none claims it evidences performance.