Understand what an AI agent can reach before you trust what it can do
ECZ-ID Agent Trust Pro helps developers, AppSec teams, platform engineers and security reviewers understand AI-agent authority, connected tools and meaningful change.
It extends the free ECZ-ID Agent Trust Community capability with deeper authority mapping, capability-chain review, change history, MCP relationship review and local policy.
Community remains free. Pro unlocks the additional Agent Trust capabilities while a valid Pro entitlement is active.
Agent security is more than the model
An AI agent can potentially edit files, execute commands, call APIs, use MCP tools, trigger workflows and perform external actions.
A useful trust review therefore needs to answer:
- What can the agent reach?
- Which tools can it use?
- Which permissions matter?
- Which capabilities become significant when chained?
- Which MCP surfaces extend its authority?
- What changed since the last review?
- Is accountable public proof available?
ECZ-ID Agent Trust Pro is designed to make those questions easier to answer inside the developer workflow.
Agent Trust Pro capabilities
Authority Graph
Map relationships between an agent, its tools, MCP connections and other reachable capability surfaces.
Authority Epochs
Capture meaningful authority state over time so later changes can be compared with an earlier known state.
Reachability analysis
Review the surfaces an agent can potentially reach through its declared configuration and connected tools.
Dangerous capability-chain review
Identify combinations of capabilities that can become more consequential when used together.
Destructive-capability analysis
Surface declared capabilities associated with deletion, overwrite, execution, deployment and other high-impact actions.
Credential-exposure review
Identify credential-shaped environment-key exposure without requiring secret values to leave the normal local inspection boundary.
External-action review
Make consequential external actions and integrations easier to identify during review.
Reciprocal MCP checking
Review relevant MCP relationships so an agent is not evaluated in isolation from the tools extending its authority.
Resolver and KYA-aware review
Surface ECZ-ID Resolver and Agent Credential / Know-Your-Agent proof routes where available.
Absence of public proof is not itself a safety verdict.
Custom local policy and richer history
Apply your own review policy and retain useful context around authority changes, findings and remediation.
Local-first by design
ECZ-ID Agent Trust is designed around local-first inspection.
For normal local analysis:
- source code does not need to be uploaded,
- prompts do not need to be uploaded,
- secret values do not need to be uploaded,
- local policy remains under customer control,
Public ECZ-ID evidence, where available, is checked separately through the ECZ-ID Resolver.
Community vs Pro
Agent Trust Community — Free
Designed for fast local AI-agent discovery and baseline review, including:
- agent/configuration discovery,
- dependency inventory,
- Capability X-Ray,
- declared permissions,
- environment key-name exposure,
- workspace MCP relationships,
- baseline and change detection,
- Resolver/operator lookup,
- supported
ecz-agent.json validation and scaffolding,
Agent Trust Pro — £12.99/month or £119/year
Designed for deeper authority review and local policy, including:
- Authority Graph,
- Authority Epochs,
- reachability analysis,
- destructive-capability analysis,
- dangerous capability chains,
- credential-exposure review,
- external-action review,
- reciprocal MCP checking,
- richer authority drift/history,
- custom local policy,
- remediation history,
Annual billing saves £36.88 — approximately 24% compared with twelve monthly payments.
Built for
- AI-agent developers,
- AppSec teams,
- product-security teams,
- platform engineers,
- AI infrastructure teams,
- enterprise architects,
- SaaS companies adding agentic capabilities,
- organisations evaluating third-party agents,
Important trust boundary
ECZ-ID Agent Trust Pro is developer security and authority-review software.
It does not certify an AI agent as safe, approved, compliant or suitable for every environment.
It does not replace secure engineering, access control, sandboxing, human review or organisational security policy.
Evidence informs the decision. Local policy decides whether reliance is appropriate. Re-check when agent authority changes.
Product specifications
Product: ECZ-ID Agent Trust Pro
Product family: ECZ-ID Developer Trust
Product class: Developer security software
Primary subject: AI agents
Primary environment: Visual Studio Code / supported compatible distribution environments
Delivery: Digital software subscription
Plan: Pro
Billing: Monthly or annual recurring subscription
Monthly price: £12.99 GBP
Annual price: £119 GBP
Community capability: Available separately at no charge
Physical shipment: None
Public proof surface: ECZ-ID Resolver
Vendor: ECZ-ID
Publisher: EcoCitizenz Ltd
