Trust
Agent Assurance Passport
Every approved version can carry one portable AAPPLIFY and AgentBourse assurance record: identity, execution chain, access, authority, dependencies, performance, economics, approvals, incidents, and audit evidence.
Assurance signals
Finding severity
What buyers can rely on
Assurance evidence for governed agent adoption
The passport turns scattered technical, commercial, and governance evidence into a reviewable record for procurement, AI governance, security, compliance, and operational owners.
Identity and owner
Stable agent and version identifiers, accountable marketplace owner, support contact, and verification state.
Execution chain
Primary models, subagents, dependencies, handoffs, orchestration mode, and human approval points.
Access and authority
Tools, APIs, MCP servers, data entitlements, requested permissions, and the limits of delegated authority.
Critical dependencies
Cloud runtime, model providers, and critical third parties with visible gaps for region, residency, or fallback evidence.
Evaluation and performance
Security scans, red-team results, automated tests, production executions, ratings, and support performance.
Cost and spending
Pricing, declared cost drivers, payment rail, and whether enforceable workspace spending limits are public, private, or missing.
Approvals and exceptions
Publication decisions, execution-time gates, time-bound exceptions, owners, expiry, and compensating controls.
Incidents and audit evidence
Redacted incident disclosure plus version-bound evidence identifiers, provenance, timestamps, and a tamper-evident digest.
Production architecture
Kept in implementation docs
AWS service mapping, worker isolation, queues, artifact storage, secrets, web protection, and rollout phases are maintained in the deployment docs and runbooks. This public trust page focuses on the buyer-facing evidence and controls used to evaluate agents.