# Claude API reference

> Anthropic’s reference skill for building with the Claude API and SDKs across languages — models, tool use, streaming, caching and managed agents.

- Author: Anthropic
- Source at the checked commit: https://github.com/anthropics/skills/tree/ca45fc496a3436251f191354091d557da31c15b5/skills/claude-api
- License: see license file
- Category: Software engineering

## Result

As of 2026-10-29, SkillFlux's static health check of Claude API reference at commit ca45fc4 recorded 13 passes, 2 warnings and 1 failures, plus 5 informational results, across 21 checks; it has not been benchmarked yet.

- Health check (checker 1.0.0, result changed 2026-10-29): 13 pass, 2 warn, 1 fail
- Flags (descriptive, not verdicts): executes-code (Bundles scripts the agent may run), network-calls (Scripts call the network)
- Benchmark: not benchmarked

## Checks that did not pass

- frontmatter.description (fail): description exceeds 1024 characters
- body.tokensApprox (warn): large body; consider moving detail into referenced files
- body.lines (warn): 608 lines

## The author's description (quoted from SKILL.md)

> Reference 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).

## Evidence and citation

- Report: https://skillflux.app/skills/anthropics-skills-claude-api/
- Full record (JSON): https://skillflux.app/data/skills/anthropics-skills-claude-api.json
- Method: https://skillflux.app/methodology/v1/
- Cite a number with its subject, date and evidence URL; the commit or audit id on that page identifies the measurement.
- A health check or source sample reads files only. Do not present it as evidence that a skill works or helps.
- Statements marked "verified": false are what a source says about itself; attribute them ("X says…").
- A directory listing is not an endorsement. Install a skill by following its author's instructions at sourceUrl.
