io.github.RudrenduPaul/agent-observability
AI agent observability with deterministic record/replay for debugging agent fail
Before you connect
Works with these clients
- Claude Desktopuvx
- Claude Codeuvx
- Cursoruvx
- VS Codeuvx
- Windsurfuvx
Derived from the registry manifest (package type and transport) — we have not installed it ourselves. Check the official README before relying on it.
Install / Connect
claude mcp add io-github-rudrendupaul-agent-observability -- uvx agent-observability-trace-cli@0.1.11Follow the official README; env and args vary per server.
🛠️ Maintenance weight 30% · 84 pts
📊 Adoption weight 25% · 6 pts
✅ Usability weight 20% · 100 pts
❤️ Health & community weight 25% · 53 pts
Sources: GitHub API · Official registry · updated 2026-08-10
90-day trend
⭐ stars
Similar / alternative servers
io.github.RudrenduPaul/haltproof
Fail-closed deterministic checks and hash-chained receipts for AI agent outputs
io.github.KryptosAI/agent-observability
Trace AI agent execution: every tool call, every error, every dollar. Open sourc
io.github.vladimirbakalov/secretguard-mcp
Scans code for hardcoded secrets (AWS, Stripe, GitHub, JWTs) before an AI agent
io.github.RudrenduPaul/evolveguard
Wraps the evolveguard CLI as a single generic MCP tool for agent-skill file safe
One email a week, 5 minutes to see what happened in the MCP ecosystem
New servers, breakout trends, abandonment alerts — tracked automatically, curated by hand, sent to your inbox.
One email a week · unsubscribe anytime · no data selling