A respectful, evidence-first read of Chevron's public web surface and public reporting: what an ESG analyst or an AI agent can and cannot verify today, and the one layer that would change it. Baseline credited first; then four findings, each reproducible from a URL and an HTTP status. Third-party research is cited as that party's claim, not ours.
Chevron's carbon and lower-carbon programs are real, long-running, and disclosed: this brief is about verifiability, not effort:
The gap is not that Chevron does nothing. It is that its most-scrutinized climate claim, the retired credit, has no proof-of-work behind it that a third party can independently trace and verify.
No jargon. Each fix below does one job: turn the climate work Chevron already does and already pays for into something an analyst, a buyer, or a regulator can independently verify. Here is what each is worth.
Right now a retired credit is just a registry serial number anyone can attack as low-integrity; bind it to proven, photographed field work and the money already spent on offsets becomes a defensible asset instead of a reputational risk.
Today the site turns every research bot away at the door, so the answer an assistant gives about Chevron is written by whoever criticized it loudest; open the door and Chevron's own record is what analysts and buyers read.
The plugging and clean-up work is already mandatory and already recorded internally; a per-site proof receipt turns that spend into credibility a community, investor, or regulator can confirm site by site.
The New Energies and methane commitments live only as prose and PDF, so an AI summarizing Chevron has to guess; make them machine-legible and the story gets represented accurately wherever agents repeat it.
Chevron retired roughly 5.8 million carbon credits across 2020–2022 through the major voluntary registries. Independent research (Corporate Accountability, reported via Carbon Herald) has publicly claimed that the large majority of Chevron's offsets are low-integrity. Whatever the truth of any single credit, the structural problem is the same: a retired credit is a registry serial number, a claim, with no traceable, independently-verifiable field work behind it. An ESG analyst asked "what real, proven work backs this retirement?" has no object to point to.
soil_sampling 720 · pfas_screen 123 · ambient_air_survey 157 · biodiversity_survey 178 · soil_carbon_baseline 129. Re-verify: curl -s https://vealth.net/labor/statschevron.com sits behind Akamai bot management that returns HTTP 403 to any non-browser user-agent: a plain curl request and a no-User-Agent request both get 403, while a full browser UA is served normally. Its robots.txt names only social crawlers (facebookexternalhit, Twitterbot, LinkedInBot) with no rules for AI crawlers at all. So an AI research agent, an ESG bot, or an A2A/MCP (the interface an AI assistant calls to use a service correctly) client is blocked at the door. With a browser UA, every agent well-known path (/llms.txt, the file AI assistants read first, /.well-known/agent-card.json, /.well-known/ai-plugin.json) is a true 404.
llms.txt + agent card the way a peer (Edison) already invites them.curl -s -o /dev/null -w '%{http_code}' -A 'curl/8.0' https://www.chevron.com/ → 403; with a browser UA → 200; curl -s https://www.chevron.com/robots.txt shows only social-bot rules.Plugging & abandonment, site remediation, and methane-leak repair are exactly the survey-a-site → do-a-bounded-task → record-it lifecycle EcoWealth runs. IIJA orphan-well programs even require quantified methane rates, a natural fit for a proof + tCO2e receipt. Today the completion of that work lives in internal and regulatory filings; nothing public lets a community, an investor, or a regulator verify that a given site was actually remediated to spec.
curl -s https://vealth.net/.well-known/x402 | grep -E 'work-packet|work-tco2e|retire'The homepage (Sitecore CMS) carries 0 application/ld+json blocks. The New Energies story, the methane commitments, and the offsets explainer are delivered as prose and PDF. An AI agent asked to assess Chevron's lower-carbon commitments against outcomes has only marketing text to parse, assuming it isn't 403'd first.
curl -s -A 'Mozilla/5.0 ... Chrome/126' https://www.chevron.com/ | grep -c 'application/ld+json' → 0.A legacy well-site remediation or restoration task becomes an EWP proof-verified work packet: bounded scope, claimed by a remediation crew, closed with photo + GPS + signature proof scored by a deterministic proof-checker, settled to a permanent receipt on Base, and where it generates a carbon outcome, the retirement settles on-chain bound to that packet via the live Klima rail. The result is the thing Chevron's offsets don't have today: a credit an ESG analyst can trace back to real, proven ground.
Live proof it's not slides: EWP contract 0x76c17C…A14B on Base mainnet (source-verified, first settled packet workId 14); the priced tool + retire menu at vealth.net/.well-known/x402; and the end-to-end building-agent golden path.
| Step | What ships | Who owns it |
|---|---|---|
| 0 · Prove it (today) | One agent pays $0.10 to work-packet (or the retire-attach lane); one funded remediation packet + bound on-chain retirement settles. | EcoWealth rail, live now |
| 1 · Machine-legible surface | Allow known-good AI agents at the edge; publish llms.txt + agent card + JSON-LD on New Energies / methane pages + security.txt. | Chevron web/security team (drafts provided) |
| 2 · Proof pilot | 10 remediation packets across one legacy field, each closed with proof + a bound retirement receipt; verified per-site history. | Joint, bounded scope |
| 3 · Standing rail | Retirements bound to proof-verified work as the default; per-site remediation receipts as the offset-integrity backbone. | Operator decision |
Two figures, both invisible against a supermajor's books:
work-packet endpoint (or the retire-attach lane for a real Base retirement) and gets a bounded, proof-framed result back, keyless. Dry-run client: vealth.net/recipes.html.