15,692 mods in this category, of every kind an
agent can take. Each one carries what it costs per session, what the
scan found, and whether it is the original.
Use when Codex must directly operate or troubleshoot the installed grok-plugin-codex MCP capability surface, including CLI/model checks, literal prompt runs or continuations, session inspection or export, and known background-job control. Do not use for bounded repository delegation, implementation, review…
MCP server for RubyHash (rubyhash.dev): diff two Ruby hashes from failing Minitest/RSpec output and convert Ruby hash literals to sorted JSON. Runs entirely locally, nothing is uploaded. Runs locally from the rubyhash-mcp npm package.
★not rated 0 2mo agoA
tokens not measured
originalMIT
Diagnose why a Tdarr transcode job failed from its job report. Use when the user asks why a particular Tdarr job or transcode failed, supplies a Tdarr JobId, or wants to inspect current failed transcodes and choose one to investigate.
Use when building, automating, or debugging Cisco Packet Tracer networks through the packet-tracer MCP (the pt tools) — planning topologies from natural language, live-deploying into a running Packet Tracer, adding devices / links / modules, applying ACL / NAT, exporting configs, or writing raw Script-Engine JS via…
MCP server that gives coding agents accurate, version-to-version breaking-change briefings for any npm package — so migrations stop being hallucinated. Runs locally from the breaking-changes-mcp npm package.
MCP server that grounds coding agents on the dependency versions ACTUALLY installed in the current project — read from the lockfile and nodemodules — so they stop coding against a blend of every version. Runs locally from the installed-version-mcp npm package.
★not rated 0 2mo agoA
tokens not measured
originalMIT
Use when debugging a running application whose behavior you can reproduce — instrument the app with logsafe structured logging, reproduce the bug, and read the session narrow-first (errors → trace → widen). Works via logsafe MCP tools when connected, or plain HTTP/curl.
When the user says “fix this”, “what is wrong”, or “debug this”, use visible terminal/editor context to create a structured debugging command for Codra / Codex / OpenCode.
Use when the user reports a 3D model or scene looking wrong (black/invisible/gigantic/untextured model, holes, weird shading), asks to check or QA a 3D file (GLB, GLTF, OBJ, PLY, STL, .splat), works with Gaussian splats, or generates 3D assets with AI and wants them verified. Requires the 3d-eyes MCP server.
Diagnose runtime bugs by connecting source-code log statements to Graylog events. Use when a user reports missing results, failed callbacks, stuck async jobs, inconsistent data, or asks to trace a code path through logs. Also use when code inspection reveals relevant log.info, log.warn, log.error, logger, tracing, or…
Live runtime inspection for any Python app — MCP bridge that lets Claude Code reach into a running Python process to eval expressions, inspect objects, read source, follow logs, and capture screenshots.