kromatv/kroma / .claude/skills/reference-first
reference-first
VerifiedGround platform and integration code in documented facts instead of memory. Before writing code that depends on how an external platform, SDK, protocol or vendor behaves (key names, manifest fields, event payloads, threading rules, supported formats, store requirements), read the project's reference docs for that area; when the fact is missing, research it from the official source, cite it, and add it to the reference in the same change. Use when implementing or debugging native, OS, device, SDK or third-party integration code, or when a behaviour "should" work a certain way. Triggers - "how does the platform handle", "what key name", "is this supported on", "according to the docs", "port to a new platform", "native module", "why does the TV".
Reference first
A platform behaviour written from memory is a bug with a delay. The fix is cheap: look it up once, write it down with its source, and let every later change read the written version.
Before writing the code
- Find the project's reference for the area (a
docs/reference/directory, a packageCLAUDE.mdthat maps code to reference files, or the architecture docs). Read the section the change depends on. - If the fact is there, follow it. If the code disagrees with it, the code is the suspect.
- If the fact is missing or vague, research it before coding:
- the vendor's official documentation and API reference;
- the SDK's own headers and sources (installed SDKs ship them);
- a mature project already shipping on that platform, read at its source.
- Write what you found into the reference in the same change: the fact, a link or a file and line, how it is known (documented, inferred from what, or verified on which device), and which versions it applies to.
While debugging
- Check the reference before guessing at a cause. Many "mystery" bugs are a rule the platform documents.
- When hardware contradicts the documentation, trust the hardware, record both, and say which device showed it.
- A fix that came from an experiment still gets a reference entry, so the next change does not rediscover it.
What not to do
- Do not state an assumption as a fact, in code or in docs. Mark it inferred, or leave it as an open question.
- Do not copy whole vendor pages into the repository. Record the facts the code depends on, with a link to the page.
- Do not cite from memory. A citation you did not open is not a citation.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization. | Stars: 465k | Updated | Claude CodeCodex | Verified | Install |
| academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file). | Stars: 196k | Updated | Claude CodeCodex | Unverified | Install |
| doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |