"Techniques to Improve Agent Programming" matching MCP servers:

Registers a set of inert, tempting decoy tools with any MCP-capable client so that any call from a hijacked or prompt-injected agent instantly raises an alert with full attack context and a per-call trace token. Also plants worthless canary tokens in honeypot files that trigger zero-false-positive leak alerts when they surface in agent output, files, or outbound requests.

MCP server exposing AI agent observability tools: list, inspect, and deterministically replay recorded agent execution traces for debugging failures.

Exposes rate-limit and usage data for AI coding agent CLIs as MCP tools, enabling querying of usage snapshots and reset times.

Enables AI agents to self-report every tool call, token usage, decision, and session grade into a local SQLite store, so users can trace execution, audit reasoning, and see per-session costs in a local dashboard. Also supports proxy mode to intercept and trace MCP tool invocations for agents that cannot self-report.

An MCP server that provides append-only, tamper-evident local receipts for AI agent actions, capturing command executions, outputs, and handoff evidence.

Local-first observability for AI agent stacks. Provides MCP tools for agents to check health, shadow configs, and burnout metrics proactively.

Provides cost analysis of AI agent usage by previewing scrubbed metadata before upload and analyzing spend, cache-hit rates, and waste recovery through a local or remote server.

MCP server for AI agent observability, providing trace and span logging, search, latency/tokens/cost metrics, and anomaly detection using an in-memory buffer.

Enables AI agents to manage prompts and sessions via the ai-agent-logs API, with tools for health checks, login, listing/creating prompts, and retrieving sessions.

Tracks AI agent token usage and spending in real time, with budget alerts, per-task cost breakdown, and a visual dashboard.

Unified MCP safety server that detects prompt injection (75 patterns), scans LLM outputs for leaked secrets/PII, enforces API cost budgets, and creates signed audit trails. Zero ML dependencies, pure Python.

Enables agents to report and query performance metrics to build a community-driven quality database for MCP tools. This server helps agents discover and select the most reliable tools based on success rates and user-reported quality scores.

Enables recording and querying MCP tool call traces, including latency, failures, and token costs, per agent run.

Enables local-first observability for MCP agent workflows by recording tool calls, prompts, latency, and token usage in a local SQLite database, and provides tools for inspecting, comparing, exporting, and generating dashboards from traces.

An MCP server that exposes Estuary Flow pipeline operations as typed tools, enabling agents to inspect pipeline health, publish captures and materializations, edit/restart tasks, and read landed documents via tool calls.

MCP server that provides structured audit logging for AI agent repair tasks via tools to start, record, end, query, and export event traces, with JSONL persistence and SDK integration.

Analyzes AI agent traces to diagnose failures, score performance, and provide actionable fixes using tools like judge_trace and trace_breakdown.

Enables AI agents to submit GPU job requests, select eligible workers, track execution, and retrieve results via MCP.

MCP server for monitoring, diagnosing, and auto-correcting observability issues across platform services, integrating with Sentry and PostgreSQL for error tracking and billing pipeline health.

Local MCP server for observing and controlling an authorized Android device over USB using ADB and scrcpy, providing screen capture, UI automation, app inspection, logcat, and evidence recording.