Agent Replay Debugger Mcp
Agent Replay Debugger MCP is a MCP server that records agent steps and replays runs deterministically for debugging.
About
Agent Replay Debugger MCP gives developers a step debugger for LLM agents: it records each step in a run and supports deterministic replay so you can see exactly which tool calls, decisions, and contexts led to a bad outcome. When you ship agent workflows in Claude Code, Cursor, or headless automation, intermittent failures are painful because chat logs alone rarely capture full state; this MCP server targets that gap with explicit record-and-replay tooling at the protocol layer. Use it during ship while you stress-test multi-step flows, then lean on it in operate when users report wrong answers or unsafe tool use. Version 1.0.1 is distributed via PyPI for stdio MCP clients; connect it alongside your other CSOAI guardrails when you need forensic detail rather than another policy gate. It does not replace unit tests or eval harnesses—it makes individual production-like runs reproducible so you can patch prompts, retrieval, or tool permissions with evidence.
- Records every agent step for later inspection
- Deterministic replay of recorded agent trajectories
- Step-debugger mental model for autonomous agents
- PyPI stdio package agent-replay-debugger-mcp v1.0.1
- MCP 2025-12-11 server manifest (stdio transport)
Agent Replay Debugger Mcp by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
claude mcp add agent-replay-debugger-mcp -- uvx agent-replay-debugger-mcpAdd your badge
Show developers this MCP server is listed on Skillselion. Paste this into your README.
| Package | agent-replay-debugger-mcp |
|---|---|
| Transport | STDIO |
| Auth | None |
What it does
Record each agent step and replay runs deterministically when an autonomous workflow misbehaves in production or staging.
Who is it for?
Best when you're debugging non-deterministic multi-step agents and need recording and replay without building a custom observability pipeline.
Skip if: Simple one-shot completions with no tools, or teams already satisfied with full OpenTelemetry traces across every hop.
What you get
After you connect this server, recorded runs can be replayed step-by-step so you can fix prompts, tools, or data with a reproducible trail.
- Recorded agent step histories accessible via MCP
- Deterministic replay sessions for investigation
- Clearer root-cause data for prompt and tool fixes
By the numbers
- Server version 1.0.1
- 1 PyPI stdio package
- stdio transport only in published manifest (no remote listed)
Recommended MCP Servers
0piFree, 2hr expiry agent storage to cache contexts, bridge workflows, or dump data.
100Hires AI ATS & Recruitment SoftwareOfficial 100Hires MCP: AI ATS & Recruitment Software for candidates, jobs, applications, interviews.123elec McpInterface MCP officielle du site marchand 123elec, catalogue, stock, panier, commandes.
1stayHotel booking MCP server.2
2D Assets MCPMCP server for generating advanced mock 2D PNG assets for game prototypes.43D MeshWeaverMesh spatial optimizer reducing geometric overhead in 3D assets.How it compares
Agent trajectory replay MCP, not a static code linter or prompt-injection firewall.
FAQ
Who is agent-replay-debugger-mcp for?
Developers and operators running tool-using agents who need to record and deterministically replay sessions when behavior goes wrong.
When should I use agent-replay-debugger-mcp?
Use it during agent QA and whenever you must investigate a specific failed or surprising run in staging or production.
How do I add agent-replay-debugger-mcp to my agent?
Install the PyPI package agent-replay-debugger-mcp and register it as a stdio MCP server in your Claude or Cursor MCP config per the server manifest.