#agent-observability

Inspect Codex / Claude logs, verify metering samples and re-count supported billing exports. Free local tools; bounded $10 diagnosis and $20 QA.

Security-First Agentic IDE Coding Harness

Local replay debugger for Browser Use failures with screenshots, model I/O, failed-step timelines, and public-safe HTML exports.

git for running agents — inspect, replay, share, fork AI coding sessions

Python SDK + CLI + MCP server for AgentClip — the open-source tool that turns AI agent runs into shareable URLs.

Guarded CLI, TypeScript client, and MCP server for JustHandled x402 agent utilities

Local-first AgentOps flight recorder and capability firewall for MCP-based coding agents with OpenTelemetry traces.

Just a lightweight lens for agent data and metrics. Built to scale past Claude when needed.

Failure diagnosis and replay for coding agents.

OpenTelemetry observability and privacy-aware failure replay for production AI agents.

Local observability and session state tracking for coding agents like Claude Code, Codex, and Qwen.

Replayable agent evidence, conflict-safe memory, and source-bound multi-agent research on flop.

Test agent behaviour like code: record a run, replay it deterministically, assert what it did, diff what changed.