Skill UI

Implements an ultra-compressed communication mode by drastically reducing token count while maintaining high technical accuracy. This mode strips filler words, articles, and pleasantries, allowing users to communicate complex technical concepts in the most concise way possible. Ideal for debugging, generating logs, or highly efficient development conversations. Supports multiple intensity levels, including classical Chinese modes.

This utility compresses verbose natural language notes, memories, and documents (like markdown files) into a highly dense format. By eliminating filler words, redundant phrases, and grammatical fluff, it significantly reduces the token count for large inputs. Crucially, it strictly preserves all technical content, including code blocks, inline code, URLs, file paths, and commands, ensuring technical accuracy while boosting AI processing efficiency.

This tool generates highly standardized and ultra-compressed commit messages following the Conventional Commits specification. It ensures that commits are concise, professional, and adhere to strict formats (e.g., type(scope): summary). It intelligently generates a body only when the context (like breaking changes or complex reasoning) requires it, significantly reducing noise and maintaining a clean, clear project history.

Streamlines the process of providing code review feedback by enforcing a highly structured, ultra-compressed format. It guides users to cut through noise and deliver actionable, single-line comments (Location: Problem. Fix.) while maintaining critical context like severity and technical rationale. Ideal for team settings prioritizing efficiency and high code quality.

This utility compresses natural language documents (notes, preferences, memory files) into a highly condensed, 'caveman-speak' format. Its primary function is drastically reducing input tokens for efficiency while meticulously preserving all technical substance, including code blocks, URLs, file paths, and markdown structure. Ideal for optimizing memory consumption when interacting with AI models.

A comprehensive reference guide and toolkit for stylistic text compression. It provides multiple modes (Lite, Full, Ultra) and specialized skills (e.g., commit messages, review comments) to transform verbose, chatty text into extremely concise, professional prose. Ideal for developers needing to write terse commit logs, academic summaries, or simply minimize token usage.

Cavecrew provides a set of specialized subagents (investigator, builder, reviewer) designed for efficient code development delegation. Unlike vanilla agents, Cavecrew compresses output results, drastically reducing the token usage injected back into the main context. This allows users to maintain context over much longer, complex coding sessions, effectively preventing context exhaustion during large-scale refactoring, debugging, or multi-step feature development.

This utility displays the precise, real-time token usage and estimated cost savings for the current session. It reads directly from the internal session log, ensuring objective and accurate data without requiring AI computation. Triggered by /caveman-stats, it provides immediate performance metrics to monitor API costs and usage patterns.

This skill operates on two layers: the THINK layer enforces rigorous coding discipline (simplicity, surgical changes, goal-driven execution), and the SPEAK layer compresses prose output (caveman/classical modes) for maximum token efficiency. Use it when brevity is explicitly requested or when disciplined, high-precision code generation is required.

This guide outlines the comprehensive process for submitting and evaluating Computer Aided Verification (CAV) research artifacts. It details the criteria for earning three badges (Available, Functional, Reusable), the two-phase review structure (smoke-test and full review), and best practices for packaging scientific tools, including the use of containers, pinned benchmarks, resource limits, and proof witnesses. This ensures reproducibility and rigor in academic research.

A comprehensive guide for drafting a technical author response (rebuttal) for peer-reviewed academic papers (CAV). It teaches authors how to structure a rebuttal by addressing specific objections related to proof soundness, benchmark fairness, novelty, and scope, using only verifiable evidence, not future promises.

This comprehensive guide outlines the critical steps for authors preparing accepted papers for camera-ready publication in the Springer LNCS open-access format. It details the process of de-anonymizing the manuscript, completing essential Springer metadata (ORCID, author order), integrating reviewer feedback, and making all research artifacts permanent using DOIs. Following this checklist is crucial for successful publication.