Agent Output Contract Validator

Store Positioning

Store title: Agent Output Contract Validator

Short description: Validate AI-agent outputs against expected schemas, required fields, source/citation rules, and structured-output contracts.

SEO title: Agent Output Contract Validator — API, schema, and developer QA

SEO description: Validate AI-agent outputs against expected schemas, required fields, source/citation rules, and structured-output contracts. Use it to catch contract drift, schema mistakes, unsafe endpoint assumptions, and developer-tool quality issues before release.

Categories: SEO_TOOLS, DEVELOPER_TOOLS

Keywords: agent, output, contract, validator, structured data, schema, api/developer qa

Pay-Per-Event Pricing

This actor uses pay-per-event pricing. Event prices include Apify platform usage; users are not expected to pay a separate platform-usage pass-through charge for the configured pricing model.

• Tier: A1 — API/developer QA
• Primary event: operation-tested at $0.00650 base
• Default max charge: $10.00
• Store discounts: FREE/BRONZE base, SILVER discounted, GOLD deepest approved discount

Event set:

• actor-start: base $0.00500, GOLD $0.00400. Agent Output Contract Validator: charged when actor start is completed. The price includes Apify platform usage; no separate usage pass-through is intended.
• operation-tested: base $0.00650, GOLD $0.00520. Agent Output Contract Validator: charged when operation tested is completed. The price includes Apify platform usage; no separate usage pass-through is intended.
• contract-rule-checked: base $0.00390, GOLD $0.00312. Agent Output Contract Validator: charged when contract rule checked is completed. The price includes Apify platform usage; no separate usage pass-through is intended.
• report-generated: base $0.05000, GOLD $0.04000. Agent Output Contract Validator: charged when report generated is completed. The price includes Apify platform usage; no separate usage pass-through is intended.

The actor accepts actor-start before work, accepts the primary event before each dataset row, and accepts the configured report event before writing report files. If maxChargeUsd or the live PPE limit blocks a charge, the corresponding row or report is not written.

Public Task Concepts

• Validate Agent Output Contract records from a capped sample
• Find invalid Agent Output Contract values before delivery
• Check Agent Output Contract coverage against expected rules
• Prioritize Agent Output Contract validation failures by severity
• Export Agent Output Contract pass-fail rows with evidence Validate AI-agent outputs against JSON schema and citation contracts.

What It Does

Validates required fields, types, enum constraints, citation/source coverage, refusal/error shape, JSON arrays, and JSONL records.

Inputs

This actor exposes these input fields in .actor/input_schema.json:

• contractJson
• schemaJson
• outputsJson
• jsonlText
• requiredCitationFields
• strictMode
• maxRecords
• includeValidRows
• maxChargeUsd

The exact default validates a bounded public GitHub API observation for modelcontextprotocol/servers, captured at 2026-07-12T05:48:00Z, against an explicit schema and source-URL contract.

Outputs

Dataset rows contain:

• recordId
• isValid
• contractRule
• fieldPath
• expected
• actual
• errorType
• severity
• fixHint
• sourceCoverageStatus

Sample outputs in examples/output.sample.json and output.sample.json were generated from examples/input.tiny.json through actor logic.

Reports

After the configured report charge is accepted, the actor writes its Markdown, JSON, or CSV reports to the key-value store. A rejected report charge leaves those files unwritten.

Public Task Concepts

• Validate Agent Output Contract Validator before release
• Run a bounded Agent Output Contract Validator QA check
• Generate developer QA evidence for Agent Output Contract Validator

Safe use

Use small capped runs first. Network paths are bounded; review output quality and costs before increasing limits.