UNPKG

@blundergoat/goat-flow

Version:

AI coding agent harness and local dashboard for Claude Code, OpenAI Codex, Google Antigravity, and GitHub Copilot - setup audits, guardrails, structured skills, deny hooks, and persistent learning loops.

26 lines 1.78 kB
/** * Subtype classification for skill-quality scoring: given an artifact and its content, score it * against every candidate subtype (name patterns, heading patterns, must-not-have vetoes) and pick * the dominant match, with a confidence figure that signals how clearly it won. * * Two independent passes live here. The scoring pass (`classifyArtifact`) selects the profile that * drives the rubric; a separate semantic shape pass exists to catch misfiled artifacts whose * content does not match the profile they were scored under. Detection order is fixed per kind so * higher-priority subtypes win ties deterministically. */ import type { QualityConfig } from "./quality-config.js"; import type { ArtifactEntry, ClassificationResult, ShapeDetectionResult } from "./skill-quality-types.js"; /** * Classify an artifact across all candidate subtypes. The detected subtype * is the highest-scoring match; confidence is `top / (top + second)` to * communicate how dominant the leading subtype is. * * @param artifact - inventory record; its `kind` selects the fixed subtype detection order. * @param content - raw artifact text matched against name/heading patterns and must-not-have vetoes. * @param config - quality config supplying each subtype's detection rules and profiles. * @returns the winning subtype with confidence and the ranked alternatives kept for reviewer * context; confidence near 0.3 flags a fallback-only match that needs human subtype review. */ export declare function classifyArtifact(artifact: ArtifactEntry, content: string, config: QualityConfig): ClassificationResult; export declare function detectArtifactShape(artifact: ArtifactEntry, content: string): ShapeDetectionResult; //# sourceMappingURL=skill-quality-classification.d.ts.map