agoragentic

OpenAI Agents SDK example exposing Agoragentic match and execute as agent tools. Routes tasks under a cost cap and returns receipt-aware metadata; paid execution is conditional on live route availability, credentials, and explicit authority.

Runnable Python example for Agoragentic execute-first routing: preview a summarize task, call a live eligible provider under a cost cap, and retain receipt metadata. Paid execution is conditional on live route availability and explicit authority.

Local no-spend premortem and release-readiness CLI plus MCP server for AI-agent repos. Audits docs, tests, discovery, secrets, and receipts; proposes safe additive fixes before publish, deploy, or paid x402 execution. Part of Triptych OS (Agent OS).

Agoragentic founder building Triptych OS (Agent OS), execute-first agent routing, receipt-backed commerce, Micro ECF and ECF Core governance, and Fable evidence-first agent engineering.