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 1.67 kB
/** * Skill/reference artifact quality HTTP route handlers for the dashboard server. * * Backs `/api/skill-quality/inventory` (list a runner's installed skill and reference artifacts), * `/api/skill-quality` (score one artifact and compose its tip prompt), and * `/api/quality/evaluate` plus its deprecated `/api/quality/analyse` alias (score uploaded or pasted * markdown). Discovery is narrowed to the selected runner's skill tree. Oversized uploads are * rejected as 413 and all validation/scoring failures are reported as JSON; alias responses also * carry Deprecation headers. Scoring engine lives in quality/skill-quality.ts; body decoding in decoders.ts. */ import type { IncomingMessage, ServerResponse } from "node:http"; import { type DashboardRouteContext } from "./dashboard-route-types.js"; /** * Bind the skill-quality handlers to one server's request context so each closure shares the path * validator, JSON responder, and body reader. * * @param ctx - per-server dashboard route context with path validation, the body reader, and IO hooks * @returns the skill-quality, inventory, and evaluate handlers; each resolves true once it has * answered a matching request, or false to let another handler claim the URL */ export declare function createSkillQualityRouteHandlers(ctx: DashboardRouteContext): { handleSkillQualityRequest: (url: URL, res: ServerResponse) => boolean; handleSkillQualityInventoryRequest: (url: URL, res: ServerResponse) => boolean; handleQualityEvaluateRequest: (req: IncomingMessage, url: URL, res: ServerResponse) => Promise<boolean>; }; //# sourceMappingURL=dashboard-skill-quality-routes.d.ts.map