Specification
Error Essence (causal) 2.0.0
A token-budgeted, trust-labeled, CAUSAL projection of one production error, intended for an LLM context window. Replaces 1.0.0's suspectCommits/changeCorrelations/similarIssues with one ranked, entity-validated candidate-cause set, and adds a sufficiency receipt recording which hydration handles were actually pulled and what each field measured. Serialized documents must estimate at or below 500 tokens — unchanged from 1.0.0 — see https://spec.bucker.io/error-essence/2.0.0.
The document
https://spec.bucker.io/error-essence/2.0.0/schema.json — served as application/schema+json, byte-for-byte the file the repository generates, so what validates here is what the code produces.
curl -s https://spec.bucker.io/error-essence/2.0.0/schema.json| Specification identifier | https://spec.bucker.io/error-essence/2.0.0 |
|---|---|
| JSON Schema dialect | https://json-schema.org/draft/2020-12/schema |
| Root type | object |
| Required at the root | schemaVersion, issueId, shortId, projectSlug, title, culprit, exceptionType, exceptionValue, level, status, actionability, frames, hiddenFrameCount, breadcrumbs, blastRadius, release, causes, refusedCauses, sufficiency, hydration, securityNotice, estimatedTokens, generatedAt |
| Named subschemas | 0 |
| Size | 21,941 bytes |
Where this identifier appears
The normative schema for the essence format, generated from the reference implementation in shared/src/essence.ts so prose and code cannot drift.
This URL will not move
A change that would alter what an existing document means gets a new version and a new URL. This one keeps resolving, because somebody has already written it into a pipeline that checks bundles without asking us anything. Every published identifier is listed on the specification index, and the namespace itself — https://spec.bucker.io — is never used for anything a reader is expected to fetch other than these documents.