This attempt to detect infinite loops and notifies when they are found. It follows the same method of [loop-protect](https://github.com/jsbin/loop-protect), which is using a timer to indicate if a loop is running too long.
Agent Runtime Control — Loop detection, prompt caching, budget enforcement, and waste scanning for autonomous AI agents.
Runtime control plane for autonomous AI agents — guards, cost limits, loop detection, and telemetry.
Observability layer for Claude Code — live token tracking, cost analytics, quota guard, loop detection, and usage dashboard. The htop for Claude Code.
When a root node matches a metadep being added to the tree, the loop-detection link deduplication made in a lot of problems as a result of the link going to the root node from the /virtual-tree temp tree, ultimately resulting in the root node being reifie
Loop detection and model stop-loss for paid AI agents. Pre-flight judgment middleware: blocks retry storms, model escalation without evidence, objective drift, and budget breaches before they cost you.
Shared policy evaluation engine for node9 — DLP, smart rules, AST shell parsing, shields, loop detection. Pure functions, no I/O. Used by both node9-proxy and the node9 SaaS firewall.
A lightweight, zero-dependency reactive state manager with automated dependency tracking, cyclic loop detection, and atomic/computed architecture.
LangChain integration for Loret — automatic loop detection and budget guardrails for LangChain agents
PromptOps observability for OpenClaw: per-call cost tracking, runaway loop detection, circuit breaker, and budget enforcement
Secure console logging library that activates sensitive data masking, infinite loop detection, and environment-aware overrides only in production environments
Vercel AI SDK middleware for Loret — native loop detection and recovery for all AI SDK providers
Agent Runtime Control — Loop detection, prompt caching, budget enforcement, and waste scanning for autonomous AI agents.
Pretty print for singly linked lists with loop detection
Reliability layer for LLM agents: scope drift, cost circuit breaks, procedural correction memory, tool-call loop detection as event-driven decisions.
[](https://www.npmjs.com/package/@aws-sdk/middleware-recursion-detection) [ and Rust SDK (#[tool], #[rule], #[agent]).
Context-native workspace manager for AI coding agents. Cuts token consumption on polyglot monorepos.
Internal crate for kdo — tree-sitter signature extraction and token-budgeted context generation. Not intended for direct use; API may change without notice.
Internal crate for kdo — core types and errors. Not intended for direct use; API may change without notice.
Internal crate for kdo — workspace dependency graph, discovery, traversal, content hashing. Not intended for direct use; API may change without notice.
Internal crate for kdo — MCP server exposing workspace intelligence over JSON-RPC/stdio. Not intended for direct use; API may change without notice.
Internal crate for kdo — manifest parsers (Cargo, Node, Python, Anchor, Go). Not intended for direct use; API may change without notice.
Policy implementations for swink-agent
Langda monitors Ruby loop operations (each, map, select, etc.), counts their iterations, measures execution time, and logs slow loops. It helps developers quickly identify performance bottlenecks in real application code with zero configuration and no breaking changes.
Leaks loops your rpsec tests and monitors process RAM usage to detect any memory leaks. Processes with continiously growing usage could protentially have a leak in them.
A production-ready AI agent platform: multi-provider (Ollama, OpenAI, Anthropic), typed tool system with permissions, three-tier memory, structured observability, loop detection, plugin architecture, interactive REPL, and project indexing. Works out of the box with local Ollama models; drop in OpenAI or Anthropic keys to upgrade.
ranma wraps a suite of Rust crates: tao (windowing and event loop), wry (WebView), Vello+wgpu (GPU 2D graphics), muda (native menus), tray-icon (system tray), arboard (clipboard), global-hotkey, and dark-light (theme detection).
No description provided.
No description provided.
No description provided.
No description provided.
No description provided.
No description provided.
No description provided.