TimeProofs AgentReady
Pre-deployment checks for agent-facing tools

Know if your API is safe enough for AI agents before they use it.

TimeProofs AgentReady scans OpenAPI specs and MCP tool definitions locally in the browser. It returns a score, risk findings, a human report, agentready.json, and static scenario simulation output.

Scan OpenAPI Scan MCP tools Preview sample report
Runs locally No API calls No MCP execution No account required Exports reports

The problem is not only whether your API works. It is whether an agent can use it safely.

Human documentation often leaves too much unsaid for autonomous tools: when to use an endpoint, when not to use it, which values are allowed, what requires confirmation, and how to recover from errors.

Ambiguous tools

Agents may pick the wrong operation when names and descriptions are too vague.

Unbounded actions

Money, deletes, emails, exports, and publishing flows need limits and confirmation rules.

Weak recovery

Generic errors leave agents guessing instead of correcting parameters or stopping safely.

One static flow for OpenAPI, MCP, and scenarios.

Start with the artifact you already have. Get a readiness contract that humans can read and future automation can consume.

1

Upload

Choose an OpenAPI JSON/YAML file or an MCP tools JSON file. The current V1 scanner runs in the browser.

2

Review

Read the AgentReady Score, severity counts, tool findings, recommended fixes, and printable report.

3

Simulate

Use agentready.json with static scenarios to check risky task patterns before live execution.

What AgentReady checks

  • Unclear operation and tool names.
  • Missing when-to-use and when-not-to-use guidance.
  • Dangerous actions without human confirmation.
  • Irreversible operations and missing success verification.
  • Unbounded numeric parameters.
  • Missing enums for closed fields.
  • Weak error recovery for agents.
  • Sensitive data and broad permission risks.

Output

OpenAPI or MCP tools JSON
-> AgentReady analysis
-> AgentReady Score
-> risk findings
-> recommended fixes
-> human report
-> agentready.json
-> static scenario simulation
-> agentready-simulation.json

Built for builders exposing tools to agents

Useful before you publish an MCP server, expose internal APIs to an AI agent, prepare agent-facing documentation, or review a tool surface before production.

Clear boundaries

TimeProofs AgentReady is not a runtime firewall and does not guarantee that an AI agent will never fail. It identifies structural risks that may cause AI agents to misuse APIs, tools, or MCP servers.