Data Quality Gate agent setup
Product, pricing and limits · Discovery index
Use Data Quality Gate with your agent
Check every row of a lead CSV before a CRM or database write.
Five trials once per verified organization; $20 plus applicable tax for 100 completed checks. No subscription or automatic replenishment.
Read the product skill · AGENT_SETUP.md · TOOL_REFERENCE.md · codex.toml · claude.mcp.json · TOOLS.json
Setup and first task
Data Quality Gate agent setup
Reviewed documentation: 2026-11-01. Examples contain placeholders, never live credentials.
Install and authenticate
Verify your email, activate this product and create an organization key in the existing account portal. Set QUALITY_API_KEY privately in the client environment. Never paste its value in prompts, public files or shell history. Codex CLI and Claude Code must be installed and signed in with existing client accounts. Missing/empty credentials must stop setup; OAuth and machine payments are not enabled. Account portal: https://mcp.contentsweep.ai/account/sign-in?product=data-quality-gate&source=agent-skill&channel=owned-agent&asset=quality-preflight&campaign=quality-lead-preflight-60d. Owned examples and installation source: https://github.com/founder-sweep/data-quality-gate-examples. Existing n8n recipes: https://mcp.contentsweep.ai/quality/workflows.
Copy this reviewed skill bundle into .agents/skills/quality-preflight/ for Codex or .claude/skills/quality-preflight/ for Claude Code in the intended project. Copy the supporting AGENT_SETUP.md, TOOL_REFERENCE.md, TOOLS.json and client examples with it. Codex user-wide skills can instead live in ~/.agents/skills; Claude Code user-wide skills in ~/.claude/skills. Merge the Codex entry into the chosen isolated profile config.toml or reviewed ~/.codex/config.toml; merge Claude entries into project .mcp.json. Never overwrite unrelated settings. Review existing settings before merging a server entry; do not replace an entire configuration file. Restart or create a fresh client session and confirm tool discovery.
Codex configuration
[mcp_servers."quality-preflight"]
url = "https://mcp.contentsweep.ai/v1/quality/mcp/"
env_http_headers = { "X-API-Key" = "QUALITY_API_KEY" }Claude Code configuration
{
"mcpServers": {
"quality-preflight": {
"type": "http",
"url": "https://mcp.contentsweep.ai/v1/quality/mcp/",
"headers": {
"X-API-Key": "${QUALITY_API_KEY}"
}
}
}
}Synthetic first task
Use the repository-owned leads.csv and leads.contract.json. Compare the passing fixture with the duplicate-near-end and warning guides at /quality/guides. Never substitute a profile or sampled summary for the original file.
Workflow
Upload the actual original bytes; draft and review the exact contract and any baseline; obtain authorized hash approval; quote; obtain authorization to use one Quality credit; start with a stable retry key; poll within a bounded deadline. Continue downstream only on continue. Download the checked original and verify its SHA256 against the receipt before writing those exact bytes. The report is not the dataset. Hash scopes differ: contract.hash, quote.contract_hash and receipt.contract_hash identify the versioned approval decision; report.contract_hash hashes only its embedded normalized contract body. Those two hash scopes normally differ, so do not compare them directly. Verify original bytes using upload.sha256, input_fingerprint in quote/receipt/report/download metadata, and the downloaded original SHA256.
Recovery and credit effects
Warnings, blocks, errors, unknown statuses and timeouts stop downstream writes for review. Completed warnings/blocks consume one check; unsupported/incomplete/platform failures are free. Resume an uncertain request using its original quote and key; do not create a second charge. Corrected bytes require a new upload, quote and key. Originals expire after 24 hours, reports after seven days and receipts after 30 days.
Price and bounds
Five trials once per verified organization; $20 plus applicable tax for 100 completed checks. No subscription or automatic replenishment.
Flat UTF-8 CSV/JSONL, 10 MiB, 50,000 records, 100 fields. No repair, remapping or independent business-meaning verification.
Official client references: https://learn.chatgpt.com/docs/extend/mcp?surface=cli and https://code.claude.com/docs/en/mcp. Skill installation references: https://learn.chatgpt.com/docs/build-skills and https://code.claude.com/docs/en/skills.
Configuration, actual MCP execution and paid customer adoption are different outcomes. Qualification fixtures and internal accounts are not customers.
Optional account and funding workflow
Open https://mcp.contentsweep.ai/account/sign-in?product=data-quality-gate&source=agent-skill&channel=skills&asset=contentsweep-quality&campaign=agent-growth-v1. The operator verifies email and creates an organization API key. Load QUALITY_API_KEY privately; never put a key in a URL, transcript or source control.
Configure Streamable HTTP MCP at https://mcp.contentsweep.ai/v1/quality/mcp/ with X-API-Key. Initialization and tools/list also require the key; anonymous 401 means authentication is required. There is no OAuth or machine-payment flow.
Five trial checks once per verified organization. Thereafter $20 plus applicable tax buys 100 checks. A completed continue, warning or block costs one check; platform/incomplete/unsupported failures release it. No subscription or automatic replenishment. Credits are product-specific. Read quality_wallet and quality_billing_status; if funding is needed, hand the operator https://mcp.contentsweep.ai/quality/app?source=agent-skill&channel=skills&asset=contentsweep-quality&campaign=agent-growth-v1. Obtain approval for the displayed purchase; never enter card details or buy automatically.
Current tool reference
Task-to-tool reference
Generated from the unchanged product MCP declarations. Parameter types and source hash are in TOOLS.json.
upload_quality_file
Upload the ACTUAL original UTF-8 flat csv/jsonl file, at most 10 MiB/50,000 records/100 fields. Stored 24h. A profile/summary/URL/server path cannot stand in for a complete file.
- Task: Inspect the supplied original file - Prerequisites: Explicit permission to read the source; remote upload also needs permission to transmit it - Outputs: Upload identity/fingerprint or bounded inferred schema - Approval: Source-file access/upload authorization - Credits: No credits consumed - Retry: Uploads create new identities; reuse the successful upload rather than uploading again
draft_quality_contract
Save a new unapproved contract version with explicit field types, numeric locale, null decisions and stable rule IDs/actions. Supported: not_null, unique, accepted_values, null_rate, row_count, range. Review before approval.
- Task: Prepare decisions for review - Prerequisites: Owned source plus explicit target rules/decisions - Outputs: Unapproved proposal, version or sampled preview; not approval or full validation - Approval: Do not infer business meanings; obtain missing decisions - Credits: No credits consumed - Retry: Reads/previews may repeat; saved drafts/versions can create new identities
get_quality_contract
Read an owned contract's exact decisions and approval hash/version before reuse.
- Task: Read an owned result or decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Structured result; inspect disclosed coverage and errors - Approval: Read access within the user's task - Credits: No credits consumed - Retry: Read again using the same resource identity
list_quality_contracts
List owned contract identities in pages of 50; read the full contract before reviewing or approving it.
- Task: Read an owned result or decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Structured result; inspect disclosed coverage and errors - Approval: Read access within the user's task - Credits: No credits consumed - Retry: Read again using the same resource identity
delete_quality_decision
Delete an owned contract or baseline after explicit operator instruction. kind is quality_contract or quality_baseline. Pending checks using it fail safely. Does not start a check or consume credits.
- Task: Delete a saved decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Deletion status; dependent pending jobs can fail safely - Approval: Explicit deletion instruction and confirmation - Credits: No credits consumed - Retry: Read current state before retry; do not delete another version
approve_quality_contract
Approve this exact contract hash only after operator authorization and review of destination meanings, locale, nulls, actions and disclosed coverage.
- Task: Record exact reviewed approval - Prerequisites: Exact reviewed hash/fingerprint and scope, with all unresolved choices settled - Outputs: Approval bound to that exact version - Approval: Explicit authorized reviewer approval; fixture permission covers only owned fixtures - Credits: No credits consumed - Retry: Reuse only the same reviewed version and scope; edits invalidate approval
draft_quality_baseline
Draft a versioned row-count baseline from the actual owned reference file for an approved contract declaring a threshold/action. Never replaces a baseline automatically.
- Task: Prepare decisions for review - Prerequisites: Owned source plus explicit target rules/decisions - Outputs: Unapproved proposal, version or sampled preview; not approval or full validation - Approval: Do not infer business meanings; obtain missing decisions - Credits: No credits consumed - Retry: Reads/previews may repeat; saved drafts/versions can create new identities
approve_quality_baseline
Approve the reviewed exact baseline hash for its specific contract after operator authorization.
- Task: Record exact reviewed approval - Prerequisites: Exact reviewed hash/fingerprint and scope, with all unresolved choices settled - Outputs: Approval bound to that exact version - Approval: Explicit authorized reviewer approval; fixture permission covers only owned fixtures - Credits: No credits consumed - Retry: Reuse only the same reviewed version and scope; edits invalidate approval
get_quality_baseline
Read owned baseline reference-file fingerprint, row count and approval/version.
- Task: Read an owned result or decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Structured result; inspect disclosed coverage and errors - Approval: Read access within the user's task - Credits: No credits consumed - Retry: Read again using the same resource identity
quote_quality_check
Quote one entire-file check with exact approved versions and original SHA256, expiring in 10 minutes. Completed blocks/warnings cost one quality credit ($20/100); errors/platform failures are free. Does not start work.
- Task: Quote an approved complete-file evaluation - Prerequisites: Owned upload and approved matching profile/contract/baseline - Outputs: Quote, expiry, exact fingerprints and funding; drift can return review_required - Approval: Approval required before quoting; quote does not start work - Credits: No credits consumed - Retry: Reuse the original quote/key for an uncertain started request; only quote anew for genuinely new work
start_quality_check
Authorize a quoted full-file preflight and reserve one QUALITY credit under the quality cap. Reuse this quote/key on retries. Completed continue/warning/block consumes one; errors release it. Feed credits cannot fund this.
- Task: Start one authorized complete-file evaluation - Prerequisites: Unexpired funded quote and stable idempotency key - Outputs: Durable job identity and status - Approval: Explicit credit/spending authorization - Credits: Reserves one product credit; supported completion (including rejection/block/warning) consumes one; platform failures release it - Retry: Original quote AND original key; changed requests need a new quote/key; never start twice after a timeout
get_quality_check
Read an owned job and operational receipt; does not start another check. Downstream writes proceed only on continue by default.
- Task: Read an owned result or decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Structured result; inspect disclosed coverage and errors - Approval: Read access within the user's task - Credits: No credits consumed - Retry: Read again using the same resource identity
get_quality_report
Read the completed full-file report for seven days: record count, rules, redacted failure counts and bounded record references. This report is NOT the original dataset.
- Task: Read an owned result or decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Structured result; inspect disclosed coverage and errors - Approval: Read access within the user's task - Credits: No credits consumed - Retry: Read again using the same resource identity
get_checked_original_download
Get the authenticated original-data download path after continue, while the 24h source retention remains. SHA256 identifies the actual checked bytes; never forward report JSON as customer data.
- Task: Fetch the checked original - Prerequisites: Owned continue job and original still within 24-hour retention - Outputs: Authenticated download path and original SHA256; verify downloaded bytes - Approval: Only continue permits downstream writes by default; separate destination authorization - Credits: No credits consumed - Retry: Read again using the same resource identity
quality_wallet
Read separate quality credits, cap, low balance and funding link. Five trials once per verified organization; $20/100 prepaid checks, no subscription/automatic replenishment.
- Task: Read an owned result or decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Structured result; inspect disclosed coverage and errors - Approval: Read access within the user's task - Credits: No credits consumed - Retry: Read again using the same resource identity
quality_billing_status
Read payment activation state for quality. Sandbox/internal qualification does not establish live activation or customer purchases.
- Task: Read an owned result or decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Structured result; inspect disclosed coverage and errors - Approval: Read access within the user's task - Credits: No credits consumed - Retry: Read again using the same resource identity
quality_purchase_history
Read owned quality purchase snapshots, settlement mode and refund state. A pack sale is one purchase of 100 credits; sandbox purchases do not establish live revenue.
- Task: Read an owned result or decision - Prerequisites: Existing owned resource and current account authorization - Outputs: Structured result; inspect disclosed coverage and errors - Approval: Read access within the user's task - Credits: No credits consumed - Retry: Read again using the same resource identity