Agent Tool Risk Audit

Audit an agent tool scope and return an evidence-backed deliverable with findings, recommended actions, and completion metadata.

What this Actor does

• Runs the exact agent-tool-risk-audit contract against only the inputs and sources documented below.
• Emits records that are validated against the bundled dataset schema before they are written.
• Keeps source references and stable record identifiers in the output when those fields are declared.
• Provides a bounded sample path for Store inspection; the sample path does not emit a paid event.

Inputs

• buyerQuery: Buyer-safe query or scope for the records returned by this product.
• taskIntent: Stable product-specific purpose for this saved Apify task.
• sourceMode: Sample emits public-safe Agentops Deliverable Status rows. Approved live source mode keeps the same output fields and only uses owner-approved public URLs.
• outputMode: Use sample records for Apify Store QA or buyer-ready records for approved Agentops Deliverable Status delivery.
• startUrls: Public URLs to use for live Agentops Deliverable Status extraction after source-policy approval. Leave empty for sample mode.
• maxItems: Caps the number of Agentops Deliverable Status rows written to the dataset.

Sources and access boundary

• User-supplied or configured inputs declared by the Actor input schema; this package does not claim an undeclared external source.

Only public or user-authorized inputs may be processed. The Actor does not bypass authentication, paywalls, CAPTCHAs, access controls, or source terms.

Output

The default dataset exposes these Actor-relevant fields:

• deliverableName: Service deliverable name.
• scopeSummary: Fixed-scope service summary.
• evidenceItems: Public-safe evidence item labels.
• recommendedActions: Recommended service next actions.
• completedAt: Example completion timestamp.

The output link exposed by .actor/OUTPUT_SCHEMA.json points to the default dataset: {{links.apiDefaultDatasetUrl}}/items.

Sample run

{
  "buyerQuery": "Deterministic fixture Deliverable Name for Agent Tool Risk Audit",
  "taskIntent": "buyer-ready-product-run",
  "sourceMode": "sample",
  "outputMode": "sample-records",
  "startUrls": [],
  "maxItems": 1
}

Example fields from the bundled deterministic sample:

{
  "deliverableName": "Agent Tool Risk Audit sample",
  "scopeSummary": "Sample scope summary for Agent Tool Risk Audit.",
  "evidenceItems": [
    "Sample evidence item for Agent Tool Risk Audit"
  ],
  "recommendedActions": [
    "Sample recommended action for Agent Tool Risk Audit."
  ],
  "completedAt": "2026-01-15T09:00:00Z"
}

Sample values are labeled as sample data and are not claims about a live source result.

Pricing and cost control

This source package preserves the Actor's current Apify pay-per-event configuration. The authenticated current event definitions are:

• risk-finding: Charge when Agent Tool Risk Audit produces Apify actor run or buyer-defined paid event.
• dataset-processed: Base charge when Agent Tool Risk Audit writes a non-empty default dataset.
• record-saved: Charge for each buyer-visible result saved by Agent Tool Risk Audit.
• enriched-record: Charge when Agent Tool Risk Audit adds match scoring, source evidence, or enrichment to a saved result.

The per-result source event is record-saved. Production output uses Apify's charged dataset push path and checks the returned charge limit before continuing. Sample, demo, test, and quality-validation paths do not emit a paid event. Start production runs with a small maxItems and an Apify run spending limit.

Reliability and limitations

• Live results depend on the supplied source availability and the fields the source actually returns.
• Missing values remain missing, null, or warning-backed according to the schema; they are not invented.
• A live run that cannot produce a schema-valid result fails instead of publishing a generic substitute.
• Output is decision-support data and does not replace legal, financial, medical, safety, procurement, or compliance review.

Existing public Actor identity

• Owner: zentrafoundry
• Actor ID: 6H9fFmodfIOHCGOFc
• Slug: agent-tool-risk-audit
• Store URL: https://apify.com/zentrafoundry/agent-tool-risk-audit

This package updates that exact existing public Actor in place. It does not create, clone, upsert, duplicate, draft, privatize, or change pricing or visibility.