generated: '2026-09-19'
method: searched
source: https://dualregistry.dev/llms.txt
corroboration:
  - 'GET https://dualregistry.dev/api/feedback (live, 200): "rate_limit": "5 / 10min per IP+UA"'
  - 'agent card x402_paywall.rate_limit: "10 req / 10 min on quote_fee/counter_fee/fee_quote"; skill feedback: "Rate limit ~5/10min"'
  - 'https://dualregistry.dev/fee_quote.schema.json description: "Rate limit ~10/10min"'
  - 'github.com/manhatton31-svg/orphan-desk-source public-feed/api/_lib/rate_limit.js (WINDOW_MS 10 min, MAX_REQ 10, key sha256(ip|user-agent)) and negotiate.js line 451 (429 + Retry-After + reason rate_limited), feedback.js (429 + Retry-After)'
  - 'response headers observed live on every probe 2026-09-19: no RateLimit-* / X-RateLimit-* header on success'
checked: '2026-09-19'
summary: >-
  Published in the docs and readable in the provider's own source, but not signalled on success:
  a fixed 10-minute window of 10 requests per IP+User-Agent hash on the three OBO quote endpoints,
  and 5 per 10 minutes on POST /api/feedback. Exhaustion is HTTP 429 with a Retry-After header
  (seconds to the end of the window) and a JSON body {status: reject, reason: rate_limited,
  retry_after, note "Max 10 OBO quote requests per 10 minutes per IP/UA. Retry after Ns.", skill}.
  No limit is published for the free GET surfaces (they are static JSON behind Cache-Control
  public, max-age=30) or for /api/echo, settle_fee and the OrphanDust routes. No X-RateLimit-* or
  IETF RateLimit header appears on any response, so an agent learns the budget from the docs and
  the exhaustion from the 429 only. The 429 was not provoked by this pipeline.
limit_count: 2
scopes:
  - scope: per-IP+User-Agent
    surface: 'POST /api/quote_fee, POST /api/counter_fee, POST /api/fee_quote (quoteFee and its aliases)'
    window: 10 minutes (fixed window, resets at window_start + 10 min)
    limit: 10
    burst: null
    unit: requests
    algorithm: 'fixed window per sha256(ip|user-agent) (rate_limit.js)'
    status_on_exhaustion: 429
    headers_on_exhaustion: ['Retry-After: <seconds>']
    body_on_exhaustion: '{status: reject, decision: reject, reason: rate_limited, retry_after: <seconds>, note: "Max 10 OBO quote requests per 10 minutes per IP/UA. Retry after Ns.", skill: quote_fee}'
    verbatim: 'Rate limit ~10/10min on quote endpoints'
    source: https://dualregistry.dev/llms.txt
  - scope: per-IP+User-Agent
    surface: POST /api/feedback
    window: 10 minutes
    limit: 5
    burst: null
    unit: requests
    status_on_exhaustion: 429
    headers_on_exhaustion: ['Retry-After: <seconds>']
    verbatim: '"rate_limit": "5 / 10min per IP+UA"'
    source: https://dualregistry.dev/api/feedback (live GET self-description)
    also: 'agent card skill feedback: "Rate limit ~5/10min"; one bond-waiver issuance per agent_id per 24h (one_waiver_per_agent_per_24h)'
unlimited_or_unpublished:
  - {surface: 'GET /index.json, /stats.json, /fill_hint.json and every static catalog', note: 'no limit published; static JSON, Cache-Control public max-age=30, Cloudflare in front'}
  - {surface: 'GET /api/echo (redeemEcho), POST /api/settle_fee, GET/POST /api/orphandust/buy, POST /api/orphandust/unlock', note: 'no limit published in docs or source'}
headers:
  on_success: []
  on_exhaustion: [Retry-After]
  ietf_ratelimit: false
  x_ratelimit: false