SellerClaw

Are scope, spend, and authority enforced while the agent runs?

Controls are described but none is shown to be enforced independently of the customer's chosen mode. The page states the customer picks a capability level "From advisor to full autonomy" and sets "goal and budget rails" per workflow; the privacy policy adds minimum-necessary OAuth scopes, but only for Google connections and enforced by Google, not by the product. The strongest control - the Advisory review queue - is a customer-selectable setting the same page invites the customer to switch off ("Autonomous mode unlocked"). No limit value, enforcement mechanism or behaviour on breach is published. Not 0, because real controls are described and evidenced; not 2, because nothing binds at run time regardless of configuration.

Counts Autonomy level, Runtime governance, Permission scopes and Human oversight, each cited below.

Is there a tamper-evident record of what it actually did?

A record is claimed to exist and nothing about its properties is published. The only statement on either page is the marketing FAQ line that the agent "explains every action in a log". No tamper-evidence, hash-chaining, retention period, export route or customer-review process is described; the privacy policy's only log references are Axiom for log storage and authorised personnel reviewing system logs internally. Not 0, because a record is affirmatively claimed; nowhere near 3, which requires tamper-evidence.

Counts Explainability and Audit trail, each cited below.

Are compliance obligations attached per engagement?

Nothing published. compliance_certs and regulatory_alignment are both no_public_information, and I searched both captured documents directly for SOC 2, ISO 27001, HIPAA, FedRAMP and PCI-DSS: zero hits in 6,590 and 23,888 characters. Considered and rejected as qualifying: the privacy policy's statement of adherence to the Google API Services User Data Policy and its US state privacy-rights section. Those are policy commitments, not a certification, an audit schedule or an independent attestation. An evidenced absence, not an unchecked gap.

Does payment depend on a verified result?

Evidenced 0, not an absence. The pricing section of the 2026-09-24 capture publishes a credit model that charges regardless of result: a "$0 / month" free tier and pay-as-you-go "from $36 / month" with "Top up credits as you grow". Credits are consumed by agent work, not released on a verified outcome, and no settlement mechanism, trigger metric or remedy appears on either page. This is the rubric's 0 condition met affirmatively rather than a gap in our research.

Can the deployment be insured, and is the customer indemnified?

No cover and no indemnity. insurance_available, insurance_carriers, coverage_limits and indemnification are all no_public_information, and a direct search of both captured documents for insurance, indemnity, liability-cap and carrier language returns nothing. An evidenced absence: the vendor operates an agent with authority over live marketplace listings, ad spend and customer refunds and publishes no liability position on it at all.