Soon every checkout will need to know whether the buyer is human, agent, or hybrid.
Sill answers, signs, and audits — so your checkout knows who's on the other side.
FOR MERCHANTS, PLATFORMS & AGENCIES · START FREE IN DISCOVERY MODE
When a mandate needs a human, the request pauses — it never auto-approves.
High-value or ambiguous agent requests escalate to a named reviewer. Approve or reject; either way the decision is written into a signed, Merkle-chained audit record with the operator's stamp. Simulated below.
Every mandate becomes a signed, timestamped record.
The bundle captures agent identity, principal delegation, intent, the full decision trace, framework mappings, and a cryptographic anchor — each record signed and independently verifiable against Sill's public JWKS.
Every record is exportable for audit submission. Compliance remains the merchant's responsibility — Sill produces the verifiable artifacts.
How Sill governs every agent interaction: identity, intent, proof
IDENTITY
Every request to the merchant carries a signed agent card naming who the visitor is, who deployed them, and what their public key proves. Unsigned traffic gets handled by your existing fraud rules. Signed traffic enters a different flow.
INTENT
Before money moves, the agent presents a signed mandate: the SKU, the cap, the merchant, the expiry, the signature. Sill evaluates it against your policy. A mandate that fails any rule never reaches Stripe to authorize the charge. Sill evaluates and signs; it never custodies funds.
PROOF
Every evaluated mandate writes an append-only, signed, Merkle-chained audit entry — yours to export. A dispute or a compliance ask becomes a query, not a forensic project.
What Sill enforces, before money moves.
Each mandate is evaluated against six categories of policy. The categories are public; your site's specific rules and thresholds live behind authentication.
Defaults ship safe. Customization is opt-in. The DSL is documented, versioned, and testable.
Signature validity, agent card freshness, organizational provenance, key rotation. Unverified agents are denied at the threshold.
Per-agent, per-merchant, per-time-window limits. Burst windows. Anomalous concurrency. Tunable per skill.
Amount caps, SKU allowlists, currency restrictions, geographic and shipping rules, bundle constraints.
Sequence analysis, velocity changes, suspicious ordering patterns, anomaly detection across an agent's request history on your site.
Merchant-defined policy expressions in a constrained DSL. Versioned, testable, deployable from the dashboard or via API.
Every decision is logged, signed, and retained. Configurable retention class. Exportable in audit-grade formats.
Test your active policy against documented attacks.
Sill's policy is a 28-rule engine mapped to MITRE ATLAS, the OWASP LLM Top 10, and the OWASP Top 10 for Agentic Applications. Each rule is a named, versioned guardrail — and every evaluation is recorded in a signed, Merkle-chained audit record.
You see exactly which scenarios your policy catches and which slip through — before they become an incident.
One script tag, between the agent and your stack.
Sill runs at the edge, in front of your existing commerce stack. Agents arrive with signed HTTP requests; Sill identifies them, evaluates each request against your policy, and forwards approved actions to your existing processor — it never custodies funds.
Signed surfaces an arriving agent can read — and anyone can verify.
When a site is verified, Sill publishes machine-readable endpoints at the edge so agents can discover its identity, skills, and catalog. Each is signed with an ed25519 envelope, and the signing key is public — so anyone can verify a signature without trusting Sill.
GET edge.sill.so/v1/agent-card/{site_key}.jsonA signed, A2A-compatible card per verified site. Advertised skills are filtered to what the site actually exposes — no overclaim.
POST edge.sill.so/v1/mcp/{site_key}A live MCP endpoint exposing the site’s backed skills as tools. tools/call invocations are recorded to the audit log.
GET edge.sill.so/v1/catalog/{site_key}.jsonA signed ai-catalog.json. Each trust manifest carries a detached ed25519 JWS over its JCS-canonical payload, attesting identity and provenance; host identity is did:web.
GET edge.sill.so/.well-known/jwks.jsonThe signing key, published openly. Anyone can verify the card and catalog signatures with standard JWS + JCS tooling — no Sill code required.
The full pipeline — signed mandate, policy evaluation, Stripe payment authorization, signed and Merkle-chained audit record — has been validated end-to-end on the live Stripe rail in production. Sill never custodies funds: Stripe holds the card, authorizes the charge, settles to your account, and pays out. Sill issues the signed authorization and the audit record. This validates the pipeline on the live rail; it is not a claim of scaled, multi-merchant payment volume.
Sill maintains public, control-by-control mappings between its guardrails and the security frameworks below.
Sill's guardrail engine maps to the LLM Top 10's highest-impact risks — prompt injection, excessive agency, and unbounded consumption.
Mandate identity, intent verification, and scope controls map to the agentic risks that matter — goal hijack, tool misuse, identity abuse, and rogue agents.
Mandate validation maps to MITRE ATLAS evasion, exfiltration, and impact tactics, including agentic supply-chain and command-and-control techniques.
Audit envelope and mandate exports support the Measure and Manage functions for AI system accountability.
These are implementation mappings, not certifications or attestations. Merchants remain responsible for their own regulatory compliance and any accredited audit their regulators require.
Pricing — agent commerce governance, priced like software
A mandate is a signed agent request to perform a sensitive action — a checkout, refund, order lookup, or shipment update. Each mandate is evaluated against your policy and produces one audit record. When paid tiers launch, billing is per mandate evaluated, not per agent visit.
The questions we get most often, answered as plainly as we can.
Does Sill process payments?
No. Sill evaluates and signs mandates; payment authorization stays with your existing processor (Stripe or your PSP). We never touch funds.
Does this replace Stripe, Shopify, or WooCommerce?
No. Sill sits in front of your existing commerce stack. Approved actions are forwarded to your checkout, refund, or order systems unchanged. You keep the same processor, the same platform, and the same data ownership.
Can agents complete checkout automatically without my approval?
Only if you allow it. Each merchant defines what agents can do unattended (typically small repeat purchases) and what requires human review (high-value, refunds, account changes). The default policy ships conservative.
What can I do with the free Discovery plan?
Identify agent traffic in your logs, publish a read-only skill manifest so well-behaved agents discover your endpoints, and see which actions agents would request. No payment authorization, no transactional authority — useful immediately for visibility.
Do I need to support new agent protocols (A2A, AP2, MCP) yourself?
No. Sill normalizes inbound agent requests and presents them to your backend in a single format. As new protocols stabilize, we add support; your integration stays the same.
Can I block unknown or unverified agents?
Yes. The default policy declines mandates from unsigned or unverified agents. You can allowlist specific agent identities, require minimum verification levels, and rate-limit by principal.
Does Sill expose my private rules?
No. The rule categories are public so visitors and auditors understand what Sill enforces. The specific thresholds, allowlists, and policy logic for your site live behind authentication and are never returned in error responses or visible to agents.
Will this work with Shopify, WooCommerce, or a custom store?
Yes. Sill installs with a one-line script tag for any site, or a CNAME-based edge install. The underlying API works with any backend that can verify a signed mandate, so it is not tied to a specific platform.
Different question? Email [email protected] and we'll answer it directly.
Add your first website. Discovery mode is free and unlimited.
Every verified site gets a signed agent card, a live MCP endpoint, and a signed ai-catalog — all independently verifiable.