Contract-audited academic research pipeline: research → write → review → revise → finalize. 5 skills (deep-research, academic-paper, academic-paper-reviewer, academic-pipeline, sr-screener), 35 modes, 16 slash commands (/ars-*), 4 agents. Skill-only, no MCP server.
5 skills · 16 commands · 4 agents · hooks
Graft code context graph for OpenClaude — 6 MCP tools (graft_find_code, graft_find_all, graft_trace_calls, graft_file_api, graft_repo_map, graft_check_freshness) + auto hooks that inject graph context into prompts, rebuild after edits, track token savings. Requires npm install -g @nanonets/graft.
1 skill · 1 MCP server · hooks
Multi-source journal research MCP (Scopus, arXiv, OpenAlex, Semantic Scholar, Crossref, DBLP, Unpaywall) with unified dedup.
2 skills · 1 MCP server
Codebase intelligence — scan, search, analyze impact of any project directory
1 skill · 1 MCP server
PR slop detection — static AST/grep scan + LLM judgment for AI-generated low-quality code before a PR is opened. Verdicts SLOP / SUSPECT / CLEAN with P1/P2/P3 findings. Skill + /slop-check command, no MCP server.
1 skill · 1 command
Codebase Memory — high-performance codebase knowledge graph MCP server (native binary v0.10.2, 158 languages). Index repos in milliseconds, sub-ms structural queries. Tools: index_repository, search_graph, trace_path, get_code_snippet, query_graph, get_architecture, index_status, list_projects, search_code, check_index_coverage.
1 MCP server
Context7 documentation lookup for any library/framework with API key rotation. 9 MCP tools: ctx7_resolve_library, ctx7_query_docs, ctx7_add_key, ctx7_remove_key, ctx7_list_keys, ctx7_set_strategy, ctx7_set_limit, ctx7_reset_stats, ctx7_current_key.
1 MCP server
Read PDF, DOCX, Excel, and TXT files. Extract text from documents for analysis and processing.
1 MCP server
Markdown to PDF/HTML/PNG conversion. PDF engine absorbed from @amitdevx/md2pdf (Playwright/Chromium): Mermaid, KaTeX, syntax highlighting, callouts, TOC, page numbers, 19 themes (14 plugin themes and 5 built-ins), custom CSS themes, headless mode. 3 MCP tools: markdown_to_pdf, markdown_to_html, markdown_to_png.
1 MCP server
INDEX-first local document navigation — fast, zero-RAM keyword search across markdown, PDF, DOCX, XLSX without heavy vector DBs. Bundles nv-index.sh (1KB structural map) + nv-search.sh (parallel ripgrep + pdftotext + pandoc). Skill-only, no MCP server.
1 skill
MCP server for managing Word, Excel, and PowerPoint documents (docx, xlsx, pptx) — 47 tools for read, write, edit, format, and export via LibreOffice.
1 MCP server
Persistent project memory via MEMORY.md. Store, search, and incrementally update project knowledge across sessions.
1 MCP server
SciSpace access for OpenClaude — headless Playwright MCP server for literature search on scispace.com using your logged-in (premium) session. 3 tools: search_papers, scispace_login, scispace_status.
1 MCP server
Vision OCR plugin for OpenClaude — extract text from images, PDFs, and URLs using RapidOCR (ONNX, CPU)
1 MCP server
Web access for OpenClaude — web search, URL fetching, and GitHub search via local MCP server
1 MCP server
PostToolUse productivity hooks — output-distiller (collapse noisy command output) + adhd-mode (focus tracking). No MCP server.