Agent skill · software engineering · jeremylongshore
anth-rate-limits
Implement Anthropic Claude API rate limiting, backoff, and quota management.
Why this skill is useful
Adds specific implementation patterns for managing rate limits and backoff strategies for the Anthropic Claude API that the AI wouldn't reliably generate on its own.
What it needs
Requires anthropic account access. About 4k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Anthropic Rate Limits Overview The Claude API uses token-bucket rate limiting measured in three dimensions: requests per minute (RPM), input tokens per minute (ITPM), and output tokens per minute (OTPM). Limits increase automatically as you move through usage tiers. Rate Limit Dimensions Dimension Header Description ----------- -------- ------------- RPM anthropic-ratelimit-requests-limit Requests per minute ITPM anthropic-ratelimit-tokens-limit Input tokens per minute OTPM anthropic-ratelimit-tokens-limit Output tokens per minute Limits are per-organization and per-model-class. Cached input tokens do NOT count toward ITPM limits. Usage Tiers (Auto-Upgrade) Tier Monthly Spend Key Benefit ------ --------------- ------------- Tier 1 (Free) $0 Evaluation access Tier 2 $40+ Higher RPM Tier 3 $200+ Production-grade limits Tier 4 $2,000+ High-throughput access Scale Custom Custom limits via sales Check your current tier and limits at console.anthropic.com. SDK Built-In Retry Custom Rate Limiter with Header Awareness Queue-Based Throughput Control Cost-Saving: Use Batches for Bulk Work Error Handling Header Description Action -------- ------------- -------- retry-after Seconds until next request allowed Sleep this duration exactly anthropic-ratelimit-requests-remaining Requests left in window Throttle if < 5 anthropic-ratelimit-tokens-remaining Tokens left in window Reduce maxtokens if low anthropic-ratelimit-requests-reset ISO timestamp of window reset Schedule retry after this time Prerequisites Record the authorized organization/model limits, budget ceiling, retry cap, and shared limiter policy. Do not infer a production limit from a local load test. Use synthetic prompts and a sandbox workspace for experiments, with no-op downstream effects and aggregate-only telemetry. Ensure logs exclude API keys, prompts, completions, tool arguments, and user identifiers; retain only headers needed to explain throttling, request IDs, and counts. Instructions 1. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/anth-rate-limits