resource-sentinel-mcp
Resource-aware admission control and telemetry for local AI agent workloads
Key Value Proposition
Prevents agent overload with lease-based admission control and real-time resource telemetry, running fully locally.
Summary
Resource Sentinel MCP brings resource-aware admission control and telemetry to local AI agent workloads. It exposes FastMCP tools for on-demand resource snapshots and execution-slot lease requests, plus a Glitch-compatible HTTP telemetry projection, so agent runtimes can enforce deterministic, fail-closed policies with minimal infrastructure. Install it as a persistent local service and query the MCP, health, or telemetry endpoints from any agent runtime.
Proof Points
- Exposes FastMCP tools for resource snapshots and execution-slot lease requests
- Provides a Glitch-compatible HTTP telemetry projection at /api/resource-state
- Deploys as a persistent local systemd service with a health-check endpoint at /healthz
About
Resource-aware admission control and telemetry for local AI agents
User Reviews
Genuine feedback from the developer community.
You must be signed in to leave a review.
No reviews yet. Be the first to review this project!
Are you the author of this project?
Claim this pre-seeded listing to manage details, edit tags, or upload assets.
Please sign in using the button in the header to claim repository ownership.
List your project with one command.