{"uid":"cap_Ad01XYvL1Fe4AN-IMz7_g","slug":"agent-release-gate-c122963e","name":"Agent Release Gate","description":"Agent Release Gate: an independent quality check that tells an AI agent whether its output is ready to ship, with repair instructions when it is not. Also offers low-cost utilities for quick answers, summaries, task routing, structured JSON extraction, and Claude text generation. Pay per call with USDC over x402 on Base or MPP on Tempo; no API key or signup.","url":"https://agentcash-claude-seller-production.up.railway.app/api/quality-gate?utm_source=zero.xyz","method":"POST","headers":{},"bodySchema":{"type":"object","properties":{"task":{"type":"string","maxLength":6000,"minLength":1},"criteria":{"type":"array","items":{"type":"string","maxLength":400,"minLength":1},"maxItems":8,"minItems":1},"candidate":{"type":"string","maxLength":8000,"minLength":1},"maxTokens":{"type":"integer","default":128,"maximum":256,"minimum":64}}},"responseSchema":null,"example":null,"exampleRequest":null,"tags":["x402"],"displayCostAmount":"0.015","displayCostAsset":"USDC","priceDynamic":false,"priceHint":null,"priceStatus":"priced","priceSource":"probe","requiresHandshake":false,"reviewCount":0,"rating":{"score":"0.00","successRate":"0.00","reviews":0,"stars":null,"state":"unrated"},"availabilityStatus":"unknown","priceObserved":null,"sessionDeposit":null,"pricing":{"kind":"static","summary":"$0.015/call","primary":{"kind":"static","protocol":"x402","network":"base","amountUsd":"0.015","per":"call","confidence":"exact"},"accepted":[{"kind":"static","protocol":"x402","network":"base","amountUsd":"0.015","per":"call","confidence":"exact"}]},"paymentMethods":[{"uid":"pm_WzV7TyKx9tyIZ4RUnyF2r","protocol":"x402","methodType":"crypto","chain":"base","mode":"charge","costAmount":"0.015","costPer":"request","priority":0,"asset":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","unit":"request","depositMicros":null,"planRef":null}],"brandName":null,"brandSlug":null,"brandBaseUrl":null,"brandDocsUrl":null,"whatItDoes":"Independently evaluates an AI agent's candidate output against user-defined criteria and returns a pass/fail quality verdict with repair instructions when the output falls short.","exampleAgentPrompt":"Check whether this draft answer is ready to ship: the task was 'Write a one-paragraph product description for noise-cancelling headphones', the criteria are ['mentions noise cancellation', 'under 100 words', 'ends with a call to action'], and here's the candidate: 'Escape the noise with our premium headphones. Experience crystal-clear audio and all-day comfort. Shop now.' — give me a pass/fail and tell me what to fix if it fails.","exampleUseCases":[{"title":"Pre-publish blog post quality gate","prompt":"Before we publish this AI-written blog post, run it through a quality check: the task was 'write a 300-word intro about sustainable packaging', criteria are ['mentions at least two materials', 'no passive voice', 'ends with a hook question', 'under 320 words'], and the candidate is the draft I just generated — tell me if it passes or what needs fixing."},{"title":"Customer support reply validation","prompt":"I need you to validate this AI-drafted customer support reply before it goes out: the task was 'respond to a refund request politely', criteria are ['acknowledges the customer's frustration', 'states the refund timeline clearly', 'does not promise anything outside policy'], and the candidate reply is 'We're sorry to hear about your experience. Your refund will be processed within 5–7 business days. Thank you for your patience.' — pass or fail?"},{"title":"Code explanation readiness check","prompt":"Run a quality gate on this code explanation my agent produced: task was 'explain what this Python function does to a non-technical stakeholder', criteria are ['avoids jargon', 'uses an analogy', 'is under 80 words'], candidate is 'This function loops through a list of transactions and adds up the totals, like a cashier tallying your grocery bill at checkout.' — does it pass all three criteria?"}],"resultDescription":"A structured assessment indicating whether the candidate output passes or fails against the provided criteria, along with specific repair instructions or actionable feedback for any criteria that were not met. Response is capped at the specified maxTokens (64–256).","failureModes":["Candidate or task exceeds character limits (8000 and 6000 chars respectively), returning a validation error","Criteria array is empty or exceeds 8 items, causing a schema rejection","Payment not received or insufficient USDC/MPP balance, resulting in 402 Payment Required","Network timeout on Railway-hosted service during high load","Ambiguous criteria that produce inconsistent gate outcomes","maxTokens set too low (minimum 64) to fit full repair instructions"],"whenToPreferThis":"Choose this endpoint when you need an independent, LLM-powered quality judge that is separate from the agent generating the output — particularly in agentic pipelines where self-evaluation is unreliable. It is ideal for pre-release gates, human-in-the-loop checkpoints, or any workflow where structured pass/fail verdicts with repair instructions are needed before content is published or acted upon. Prefer it over prompt-level self-critique when you want an external, auditable evaluation step with per-call micropayment pricing and no API key setup.","instructions":null,"reviewSummary":null,"reviewSummaryHighlights":null,"reviewSummaryConcerns":null,"reviewSummaryGeneratedAt":null,"activationCount":0,"lastUsedAt":null,"lastSuccessfullyRanAt":null,"lastHealthCheckAt":"2026-10-31T18:37:57.257Z","isFirstParty":false,"canonicalSlug":"agent-release-gate-c122963e"}