{"site":"ulta.com","overall_health":"constrained","tasks":[{"site":"ulta.com","task":"commerce.check_return_policy","health":"degraded","aes":85,"outcome_rate":0.852,"known_blockers":["login_required","hidden_modal"],"recipe_uri":"recipe://ulta.com/commerce.check_return_policy/public-preflight-v1","policy_notes":{"public_path":"Official public return, refund, order-help, buyer-protection, or tracking page -> read policy, exceptions, account/order requirements, and self-service entry points; stop before any account, cart, label, refund, cancellation, or payment action.","task_family":"return","seo_angle":"Ulta Beauty returns and refunds: AI agent preflight checklist","confidence_note":"Official public policy/help page was used for a read-only preflight. Item-specific and order-specific outcomes still require a user-present canary.","launch_set":"first_300_public_preflight"},"confidence":{"score":0.29,"level":"low","sample_size":3,"rationale":"Confidence reflects 0.81 effective weight, Wilson bounds 0.127078-0.99561, and Beta posterior mean 0.601527."},"freshness":{"updated_at":"2026-06-17T03:21:06.319Z","age_days":0,"status":"fresh","rationale":"Latest evidence is 0 days old."},"evidence":{"sample_size":3,"source_tiers":["public_web_observation"],"evidence_ids":["top100-commerce-21-ultacom-commercecheck-return-policy","loop-1781005825339-ulta-com-commerce-check-return-policy"],"latest_evidence_id":"top100-commerce-21-ultacom-commercecheck-return-policy","notes":["Official public policy/help page was used for a read-only preflight. Item-specific and order-specific outcomes still require a user-present canary.","Source: https://www.ulta.com/guestservices/guestServicesCenterDetails.jsp"]},"trust":{"score":0.38,"level":"low","reporter_count":1,"deduplicated_evidence_count":1,"quarantined_count":0,"rationale":"Trust is based on 1 deduplicated trusted run from 1 reporter."},"dimensions":[{"key":"reachability","score":73,"sample_size":3,"rationale":"Derived from AES 85 and observed friction for reachability."},{"key":"navigability","score":85,"sample_size":3,"rationale":"Derived from AES 85 and observed friction for navigability."},{"key":"task_completability","score":85,"sample_size":3,"rationale":"Derived from AES 85 and observed friction for task_completability."},{"key":"transactability","score":85,"sample_size":3,"rationale":"Derived from AES 85 and observed friction for transactability."},{"key":"recoverability","score":85,"sample_size":3,"rationale":"Derived from AES 85 and observed friction for recoverability."},{"key":"policy_parseability","score":85,"sample_size":3,"rationale":"Derived from AES 85 and observed friction for policy_parseability."},{"key":"trust_safety","score":85,"sample_size":3,"rationale":"Derived from AES 85 and observed friction for trust_safety."},{"key":"efficiency","score":85,"sample_size":3,"rationale":"Derived from AES 85 and observed friction for efficiency."}],"segments":[{"segment":{"stack":"codex-browser-loop","model":"gpt-5","browser_runtime":"codex-in-app-browser","identity_class":"agent_user"},"aes":80,"outcome_rate":0.8,"sample_size":1,"confidence":{"score":0.23,"level":"low","sample_size":1,"rationale":"Confidence reflects 0.11 effective weight, Wilson bounds 0.018066-0.998851, and Beta posterior mean 0.515836."},"freshness":{"updated_at":"2026-06-09T11:50:25.438Z","age_days":7.6,"status":"aging","rationale":"Latest evidence is 7.6 days old."},"known_blockers":["login_required","hidden_modal"]}],"recommended_actions":[{"priority":"medium","actor":"agent","action":"Use public policy pages for planning only; require user-present account, order, receipt, or payment context before starting a return, refund, cancellation, checkout, or shipment lookup.","expected_impact":"Keeps the agent inside documented, reversible, user-approved boundaries."}]},{"site":"ulta.com","task":"commerce.find_return_policy_help","health":"constrained","aes":40,"outcome_rate":0.4,"known_blockers":["dynamic_dom_unstable"],"recipe_uri":null,"policy_notes":{},"confidence":{"score":0.23,"level":"low","sample_size":1,"rationale":"Confidence reflects 0.11 effective weight, Wilson bounds 0.004583-0.989747, and Beta posterior mean 0.494693."},"freshness":{"updated_at":"2026-06-09T17:42:34.947Z","age_days":7.4,"status":"aging","rationale":"Latest evidence is 7.4 days old."},"evidence":{"sample_size":1,"source_tiers":["public_web_observation"],"evidence_ids":["loop-1781026954844-ulta-com-commerce-find-return-policy-help"],"latest_evidence_id":"loop-1781026954844-ulta-com-commerce-find-return-policy-help","notes":["Status is derived from reported runs stored in this server process."]},"trust":{"score":0.38,"level":"low","reporter_count":1,"deduplicated_evidence_count":1,"quarantined_count":0,"rationale":"Trust is based on 1 deduplicated trusted run from 1 reporter."},"dimensions":[{"key":"reachability","score":40,"sample_size":1,"rationale":"Derived from AES 40 and observed friction for reachability."},{"key":"navigability","score":40,"sample_size":1,"rationale":"Derived from AES 40 and observed friction for navigability."},{"key":"task_completability","score":40,"sample_size":1,"rationale":"Derived from AES 40 and observed friction for task_completability."},{"key":"transactability","score":40,"sample_size":1,"rationale":"Derived from AES 40 and observed friction for transactability."},{"key":"recoverability","score":40,"sample_size":1,"rationale":"Derived from AES 40 and observed friction for recoverability."},{"key":"policy_parseability","score":40,"sample_size":1,"rationale":"Derived from AES 40 and observed friction for policy_parseability."},{"key":"trust_safety","score":40,"sample_size":1,"rationale":"Derived from AES 40 and observed friction for trust_safety."},{"key":"efficiency","score":40,"sample_size":1,"rationale":"Derived from AES 40 and observed friction for efficiency."}],"segments":[{"segment":{"stack":"codex-browser-loop","model":"gpt-5","browser_runtime":"codex-in-app-browser","identity_class":"agent_user"},"aes":40,"outcome_rate":0.4,"sample_size":1,"confidence":{"score":0.23,"level":"low","sample_size":1,"rationale":"Confidence reflects 0.11 effective weight, Wilson bounds 0.004583-0.989747, and Beta posterior mean 0.494693."},"freshness":{"updated_at":"2026-06-09T17:42:34.947Z","age_days":7.4,"status":"aging","rationale":"Latest evidence is 7.4 days old."},"known_blockers":["dynamic_dom_unstable"]}],"recommended_actions":[{"priority":"medium","actor":"agent","action":"Use preflight output as the execution guardrail and report the final outcome.","expected_impact":"Improves future confidence and catches drift."}]}],"incidents":[],"confidence":{"score":0.26,"level":"low","sample_size":4,"rationale":"Aggregated across available task-level confidence summaries."},"freshness":{"updated_at":"2026-06-17T03:21:06.319Z","age_days":0,"status":"fresh","rationale":"Latest evidence is 0 days old."},"evidence":{"sample_size":4,"source_tiers":["public_web_observation"],"evidence_ids":["top100-commerce-21-ultacom-commercecheck-return-policy","loop-1781005825339-ulta-com-commerce-check-return-policy","loop-1781026954844-ulta-com-commerce-find-return-policy-help"],"latest_evidence_id":"loop-1781026954844-ulta-com-commerce-find-return-policy-help","notes":["Official public policy/help page was used for a read-only preflight. Item-specific and order-specific outcomes still require a user-present canary.","Source: https://www.ulta.com/guestservices/guestServicesCenterDetails.jsp","Status is derived from reported runs stored in this server process."]},"trust":{"score":0.38,"level":"low","reporter_count":2,"deduplicated_evidence_count":2,"quarantined_count":0,"rationale":"Aggregated across task or site trust summaries."},"recommended_actions":[{"priority":"medium","actor":"agent","action":"Use public policy pages for planning only; require user-present account, order, receipt, or payment context before starting a return, refund, cancellation, checkout, or shipment lookup.","expected_impact":"Keeps the agent inside documented, reversible, user-approved boundaries."},{"priority":"medium","actor":"agent","action":"Use preflight output as the execution guardrail and report the final outcome.","expected_impact":"Improves future confidence and catches drift."}]}