Privacy-first PDF utility (Zero-Server Architecture). Merge, split, compress, and edit PDFs 100% locally on your device. No uploads, no servers, no tracking.

14-stage Fusion Pipeline for LLM token compression — reversible compression, AST-aware code analysis, intelligent content routing. Zero LLM inference cost. MIT licensed.

Autonomously improve a real artifact (code, training recipe, agent harness, data pipeline, prompt) against an objective and an evaluator, using Hypothesis Tree Refinement (HTR) from the Arbor paper. Use this whenever someone wants to iteratively optimize something over many experiments without overfitting — e.g. "get my model's eval score up", "improve this agent/harness", "tune this pipeline", "beat the baseline on this benchmark", "run a search over approaches and keep the best", "do an MLE-bench / Kaggle-style optimization", or any long-horizon "make this artifact better and don't just memorize the dev set" task. Trigger it even when the user doesn't say "Arbor" or "hypothesis tree" but describes repeated experiment-and-evaluate loops, branching exploration of competing ideas, or worries about a dev/test gap. Runs Claude itself as the coordinator with subagent executors in isolated git worktrees; for the standalone `arbor` CLI tool see references/arbor-upstream.md.

Free open-source web software for signing PDF (alone or with others) and also organize pages, edit metadata and compress pdf

Compress LLM context to save tokens and reduce costs

Privacy-first, offline PDF editor for Android. Merge, split, compress, convert and annotate PDFs — 100% on-device, no internet required.

A collection of token-saving skills for LLM agents to compress output, trim context, and cap budget.

AI-powered PPT generation — 40,000+ style combinations, narrative-driven, design-intelligent, AI images, fully editable .pptx. Three modes: Build (default) + VI Build + FreeStyle (quick draft). 8 goal-type layouts, 35 moods, README parsing, size-aware image assignment, 3 structurally-different build.py proposals, brand compliance. Engines: Seedream, GPT Image, DALL-E, Wanx, Kimi.

Use when working with Neo4j command-line tools — neo4j-cli (modern unified

Precomp, C++ version - further compress already compressed files

Design a GitHub social-preview card (og:image, 1280x640) for a repository. Use when the user asks for a social preview, repo cover, og image, link card, opengraph image, or README hero. Four editorial moods (editorial, poster, blueprint, gallery), CJK-first typography, one self-contained HTML file, deterministic checks, crisp PNG export. No image model needed.

Think and deliver like a management consultant from McKinsey, BCG, or Bain. Use when the user wants to: (1) Structure a business problem with hypothesis-driven decomposition, (2) Run strategy analysis with professional frameworks: market sizing, competitive landscape, financial modeling, SWOT, Porter's, (3) Build MBB-quality deliverables: executive summaries, strategy deck outlines, decision memos, (4) Apply firm-specific methodology: McKinsey verdict-first, BCG framework-first, or Bain decision-first, (5) Package analysis for non-consulting audiences: investor pitches, board presentations, conference talks. Produces structured analysis and deliverable CONTENT. For visual production, hand off to a delivery skill for slides, documents, or spreadsheets.