Turns GitHub repository history into a cited maintainer skill for coding agents, providing tools to collect evidence, query the knowledge graph, and inspect bundles.
Enables Claude to use Google Gemini as a secondary AI through MCP for large-scale codebase analysis and complex reasoning tasks. Supports both Gemini Flash and Pro models with specialized functions for general queries and comprehensive code analysis.
Enables AI coding agents to query a local codebase map for session briefs, connection and blast-radius analysis, file-anchored memory with stale-note flagging, and cleanup leads. It runs locally through MCP without network access or API keys.
Enables blazingly fast file and content searching in large codebases using ripgrep, with intelligent filtering, fuzzy finding, and directory tree visualization while respecting .gitignore and avoiding common bloat directories.
Reduces Claude's context window costs by automatically summarizing inactive files to their public interfaces using AST parsing, keeping only the full contents of the currently active file.
Enables AI agents to explore, analyze, and search file systems, with tools for directory trees, file statistics, duplicate detection, code counting, and multi-criteria file search.
Provides deep git file-level forensics tools to analyze detailed file histories, changes, and patterns within git repositories, focusing on individual file analysis rather than repository-wide operations.
Enables offline inspection, structural comparison, and transactional staged editing of local copies of Omron Sysmac Studio (.smc2) projects, covering variables, POU, tasks, I/O bindings, and selected DataTypes with preview, validation, rollback, and export. Optionally integrates with a locally installed IDE on Windows for open-copy, rebuild, and guarded export operations without ever connecting to a PLC.
A validation layer for AI coding assistants that enforces explicit LLM evaluations on plans, code diffs, and tests to ensure safer and higher-quality code.
An MCP server that provides a flexible lens into directory structures and files, enabling LLM clients to efficiently navigate and understand codebases with minimal noise. It offers secure, gitignore-aware file access with tools like directory listing, file reading, and grep-like search.
Enables precise file patching using SEARCH/REPLACE blocks, with mandatory directory sandboxing, automatic Python QA (Ruff/Black/MyPy), and git versioning for rollback.
Enables intelligent handling of large files through smart chunking, search with regex support, line navigation, and streaming capabilities without loading entire files into memory.
Enables AI coding agents to extract Figma design data and convert it into Flutter widgets and screens. Supports theme setup, component analysis, asset exports, and provides Flutter implementation guidance from Figma designs.
Enables migration of test automation projects from WebDriverIO to Playwright using AST-based transformations. Provides tools for analyzing tests, converting syntax, refactoring to Page Object Model, and generating migration reports.
Enables read-only auditing of repositories by scanning actual files, flagging risky code, secrets, and untrusted platform configs, and generating Chinese-language reports for human decision-making.
Enables MCP clients such as Claude Desktop or Cursor to inspect a coffee-shop sales database schema, validate SQL, and run guarded read-only queries. It supports natural-language questions by exposing tools for schema description, SQL checking, and safe query execution.
Codebase Q&A agent with git-history awareness. Explore a repo with list_files, read_file, and search, plus git log, blame, and diff and a repo map to answer "why did this change?" questions. Runs locally as an MCP server for Claude Code and other MCP clients.
Analyzes source code structure across multiple languages using tree-sitter, extracting classes, functions, methods, and metadata with precise line numbers for efficient codebase exploration and AI-assisted development.
Enables executing arbitrary Joern/CPGQL queries against a loaded Code Property Graph, loading source code in multiple languages (PHP, Java, JavaScript, Python, C/C++) to build new CPGs, and inspecting loaded projects and the active workspace. It communicates with a running Joern server over its synchronous HTTP query endpoint, so no state polling is required.
Enables Claude to read, write, and delete files within a sandboxed base directory and to run guided code-review and documentation-generation workflows using MCP tools, resources, prompts, progress reporting, and user elicitation.
Enables Claude Web to securely inspect, search, and modify local project files through a sandboxed MCP server with atomic writes, path traversal protection, and sensitive file blocking.
A lean, local knowledge graph that joins a repo's code to its aSPARK delivery artifacts, enabling agents to trace user stories to code and assess impact of changes, served over MCP.
Enables interaction with SAP ABAP systems via ADT APIs, allowing management of ABAP objects, transport requests, and code analysis through MCP clients.
Bridges DeepSeek Harness agents to the MCP Server built into IntelliJ-based IDEs, translating the IDE's legacy HTTP+SSE transport into the stdio connection the host speaks. Exposes the IDE's own build and run, code search, refactoring, inspection, debugger, database, Git and terminal tools as callable agent tools, plus a doctor tool that diagnoses missing connections.
Enables auditing coding-agent claims against local execution traces to detect contradictions, exposing session audits, recent failures, and optional scope-alignment triage.
Enables AI agents to analyze and diff binaries by driving Ghidra headless or inside its GUI, exposing native reverse-engineering tools such as listing functions and segments, searching strings, decompiling, disassembling, and following cross-references. Ships with a devcontainer and supports local stdio or remote HTTP deployment.
Enables MCP clients to load eight on-demand coding skills and inspect project repositories through bounded, read-only map, search, read, diff, checks, and style tools. It keeps context lean by loading skill bodies only when needed and limiting repository scan and output sizes.
Sentinel-IR-MCP
MCP server for sentinel-ir:
it lets AI agents in Claude Desktop, Cursor, Windsurf or any other
Model Context Protocol client read a
compact structural IR of JavaScript, TypeScript , Python And Rust files instead of the raw source.
Enables read-only access to the canonical-vault GitHub repository through MCP, letting clients list files, retrieve file contents and history, inspect commits, and search code.
Gives OpenAI Codex behavioral history from workspace.json, exposing file fragility and co-change partners as tools to enforce safer edits and block incomplete patches.
Enables Claude Code to hand off mechanical work — file edits, code lookups, and lint triage — to a local Ollama model that reads and edits files directly, so file contents never enter Claude's context. Returns only a compact receipt, and reports success only after the change passes a deterministic syntax/project gate, rolling back and escalating on failure.
Knot is a semantic and structural codebase indexer designed for AI coding agents and developers navigating large projects. It combines vector search and graph traversal to find code by meaning, analyze impact via reverse dependencies, and explore file architectures.
Call graph MCP server for Rust, Python, TypeScript, TSX, and JavaScript. Every caller and callee carries a resolution field showing how the reference was matched to a definition, plus multi-hop impact analysis for refactoring. Local-only, single binary, no network egress. Benchmarked against grep, ripgrep, and ctags on a 120-item hand-verified golden set, including the cases where it loses.
SkillTotal is a local stdio MCP server that lets an agent check a component before installing it. It statically analyzes MCP servers, agent skills, npm/PyPI packages and repos for supply-chain risks, dangerous capabilities (shell, filesystem, network, credentials), prompt-injection surfaces like tool poisoning, and exfiltration paths. Each finding cites file, line and snippet.
Local stdio MCP server (npx -y caveat-lang mcp) with five tools that validate, check, test and explain programs in Caveat, a rules language whose decisions keep their evidence, grounds and caveats. It also finds what depends on a piece of evidence, so a withdrawn input still leaves a past decision's original basis on record.
Enables LLM agents to explore JavaScript/TypeScript repositories through parsed syntax trees by returning compact file outlines, precise symbol definitions, and local dependency graphs. It keeps access read-only and sandboxed to chosen roots while using a bounded, flood-proof stdio transport.
coldstart gives coding agents durable memory of a specific codebase, with no embeddings, no vector store, and no API key. Agents write notes after real tasks — what a file is for, how a flow spans files, which invariants hold — and those notes are surfaced automatically to a later session whose task matches, so nothing gets rediscovered from scratch. Every note is anchored to real files by content
Provides code-aware context compression by stripping comments, docstrings, and whitespace while maintaining full logic fidelity for AI agents. It features tools for architectural mapping, symbol searching, and token-budgeted multi-file reading.
Local MCP server for Windows and macOS: Claude reads your C#, Python, TypeScript and JavaScript through a syntax index, checks NuGet, PyPI and npm APIs at your pinned versions, remembers how each feature flows through your code, and edits, builds, tests and runs git. Approvals happen only on your machine; secrets are hidden.
Enables AI-driven semantic code search using Windsurf's reverse-engineered SWE-grep protocol to query local codebases with natural language. It executes local search tools like ripgrep and tree-node-cli to return relevant file paths and line ranges to MCP-compatible clients.
An MCP server that gives Claude structural understanding of a codebase—dependency graph, entry points, complexity ranking, and import cycles—without reading every file into context.
Enables coding agents to get a compact, token-budgeted map of any repository in one call, including file tree, language stats, entry points, and key symbols, so they read only what they need.
Enables efficient codebase exploration with tools to list exports, get file outlines, read targeted regions, and find references without loading entire files.
Enables coding agents to answer from what is actually true in the current Terraform working directory: the provider schemas pinned in the lock file, saved plans, redacted state, and local or installed module interfaces. It also reports what a pinned provider upgrade would break, all read-only and without ever running init in the root module.
Analyzes a local codebase and exposes its structure, per-file summaries, dependency graph, and complexity hotspots as tools for MCP clients like Claude.
Provides tools to revise and refactor code via MCP, including moving code, indenting/dedenting, outlining file structure, and optional symbol renaming.
Enables AI agents to safely read, search, and inspect a local git repository, including PR-style diffs and changed-file summaries. It also supports opt-in allowlisted test runs.
Indexes a mono-repo into a knowledge graph and provides MCP tools to query code structure—packages, components, routes, HTTP calls—without file reads or grep round-trips.