← ClaudeAtlas

honey-consolelisted

Build, change, or review the Honey LLM Proxy backend, enterprise control-plane APIs, self-hosted console frontend, ESON surfaces, security controls, or project architecture. Use for every feature, refactor, UI, API, test, and documentation task in honey-llm-proxy so work preserves the canonical Honey Console design, small-file structure, truthfulness, and three-pass verification workflow.
Green-PT/honey-llm-proxy · ★ 0 · Web & Frontend · score 67
Install: claude install-skill Green-PT/honey-llm-proxy
# Honey Console Build the secure, self-hosted Honey control plane without drifting from its established frontend or claiming controls that the backend does not enforce. ## Workflow 1. Inspect existing modules and search for equivalent functions before adding code. 2. Define or update the authenticated backend contract before implementing its frontend consumer. 3. Complete three cycles for the backend, then three for the frontend: - implement the smallest complete slice; - review security, correctness, duplication, structure, and failure states; - fix every confirmed finding and rerun relevant checks. 4. Run the complete verification suite after both parts. Do not count a review without a concrete inspection or executable check. Record each cycle in the working plan. ## Architecture - Keep one responsibility per module and one canonical helper per behavior. - Search with `rg` before creating a function, type, endpoint, component, token, or formatter. - Prefer files below 250 lines; keep authored source files at or below 400 lines. Split by responsibility before crossing the limit. Generated/vendor files are the only exception. - Keep protocol translation, streaming state, authentication, configuration, middleware, console APIs, and UI views separate. - Scope every tenant store query by organization ID. Store only hashes of random API secrets, return plaintext once, and record identity mutations in the append-only audit ledger. - Require the identity database in