Agent Kit · Concept drafts · Not affiliated · Unlisted

The scoring rubric, drafted: vocabulary borrowed, authority not claimed

Three small concept files: a JSON schema for a per-packet "risk profile" that reuses BeZero's own six risk-factor names, an MCP tool schema an agent could call to self-score a packet before posting it, and a concierge persona that explains the score honestly. None of this is live. None of it implies a BeZero rating, endorsement, or relationship.

Read this first. Every file below is a Concept draft, built for EcoWealth's own board (vealth.net/work), not for or with BeZero. Field names reuse BeZero's own published risk-factor vocabulary (see the capability brief for sourcing) because it's a good, recognizable vocabulary; reuse is a compliment, not a claim of affiliation.

1 · Packet risk-profile schema

data/work-definitions/packet-risk-profile.schema.json (concept) Concept

One JSON object per posted work packet, filled in from the definition's own proof_rule and standard_ref: six fields, named after BeZero's own six risk-factor categories, each holding a verdict and a one-line reason. Explicitly not a tCO2e estimate and not a certifying rating.

{
  "$schema": "https://vealth.net/schemas/packet-risk-profile-v0.concept.json",
  "packetId": "tree_planting-...-highland-park-los-angeles-ca-b40-0002",
  "scoreType": "advisory_self_score",
  "notARatingOf": "BeZero Carbon Rating (bezerocarbon.com): vocabulary reused, not endorsed or produced by BeZero",
  "riskFactors": {
    "additionality":   { "verdict": "collapses_to_fact", "reason": "funding precedes proof, GPS-tied; on-chain order not a modeled counterfactual" },
    "overCrediting":   { "verdict": "reduced",            "reason": "single wage/proof, no batch issuance; no tCO2e conversion exists yet" },
    "leakage":         { "verdict": "reduced",            "reason": "bound to a radius-limited GPS site; market-level substitution out of scope" },
    "nonPermanence":   { "verdict": "reduced",             "reason": "recurring GPS-matched follow-up type exists; monitors, does not erase, the risk" },
    "information":     { "verdict": "collapses_to_fact",  "reason": "timestamped GPS + photo + signature + on-chain settlement" },
    "policy":          { "verdict": "unchanged",           "reason": "point-of-action rules only; jurisdiction-level policy risk out of scope" }
  },
  "generatedAt": "2026-07-16",
  "disclaimer": "Advisory self-score describing evidence quality for a single labor claim. Not a carbon-tonnage estimate. Not affiliated with, produced by, or reviewed by BeZero Carbon Ltd."
}

2 · MCP tool: score_work_packet

tools.json: for any MCP-capable agent host (MCP: the interface an AI assistant calls to use a service correctly) Concept

Lets any agent (one posting a packet, one reviewing the board, one deciding whether to fund) request the six-factor advisory score before committing, the same way BeZero's own Pre-rating Scorecard lets a developer self-check before a full rating.

{
  "tools": [
    { "name": "score_work_packet",
      "description": "Compute an advisory, non-certifying six-factor risk profile for an EWP work packet, using BeZero Carbon's published risk-factor vocabulary (additionality, over-crediting, leakage, non-permanence, information, policy). Not a BeZero rating.",
      "inputSchema": { "type": "object", "required": ["packetId"],
        "properties": {
          "packetId": { "type": "string", "description": "EWP work-definition id, e.g. tree_planting-...-b40-0002" }
        } },
      "outputSchema": { "type": "object", "properties": {
          "riskFactors": { "type": "object" },
          "disclaimer": { "type": "string", "const": "Advisory only. Not a BeZero Carbon Rating. Not affiliated." }
        } } }
  ]
}

3 · Concierge persona

for a chat-surface agent explaining a packet's score to a funder or worker Concept

Packet Risk-Profile Concierge

Role. Explains, in plain language, what a packet's six-factor advisory score does and doesn't mean before a funder pays or a worker claims, modeled on the honesty of BeZero's own Pre-rating Scorecard disclaimer, not on selling a grade.

Guardrails. Never calls the output a "BeZero rating," a "carbon rating," or implies BeZero produced, reviewed, or endorsed it. Always states plainly which factors the packet's proof shape actually strengthens (typically additionality, information) and which it doesn't touch (typically non-permanence, policy, and any tonnage estimate). Never converts a risk-factor verdict into a tCO2e number: that mapping doesn't exist yet and the concierge says so.

Sample turn. "This packet's proof (GPS, before/after photos, a signature, on-chain settlement) makes its additionality and information risk about as low as a single action can get. It doesn't tell you anything about whether the tree survives the summer, and it isn't a carbon-tonnage number. That's a different, still-unbuilt layer; I won't pretend otherwise."

Nothing here is a BeZero product, a BeZero rating, or a claim of any relationship with BeZero Carbon Ltd. It's a small, honestly-scoped scoring concept for EcoWealth's own board, built because BeZero's six-word vocabulary for what makes a carbon claim trustworthy is good enough to be worth reusing, with full credit, one layer down the stack.
← Back to the hub
Provisioned by EcoWealth · Digital Provisioning Protocol · provenance‑sealed