Skill UI

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 skill generates comprehensive, tailored instruction files for AI coding assistants like GitHub Copilot and Claude. It supports multiple strategies (flat or nested) and is specifically designed for managing complex monorepos by creating area-scoped instructions using `applyTo` globs. Use it to standardize AI tooling usage across large codebases.

A comprehensive playbook for bypassing 401/403 Forbidden errors. Covers advanced techniques like path normalization tricks (encoding, traversal), HTTP method tampering, IP spoofing headers, and protocol downgrade attacks. Essential knowledge for penetration testing and security auditing.

This comprehensive playbook provides expert techniques for classical cipher identification and breaking. It covers fundamental methods like frequency analysis, Kasiski examination, and Index of Coincidence (IC), enabling the decryption of monoalphabetic, polyalphabetic, and transposition ciphers (Vigenere, Caesar, Playfair, etc.). It also guides the decoding of various formats (Base64, Hex, Morse).

A deep-dive playbook covering expert techniques for Path Traversal and Local File Inclusion (LFI). It details multiple bypass methods (encoding, null bytes, wrappers), target file enumeration on Linux and Windows, and advanced Remote Code Execution (RCE) escalation paths, such as log poisoning via PHP wrappers.

This guide details how to configure SMTP email delivery settings within a Webiny project using the `<Api.Mailer.Smtp />` component. It specifies the required properties (host, port, user, etc.) and emphasizes best practices, such as passing sensitive passwords via environment variables rather than hard-coding them. Ideal for developers setting up external email services.

This alias serves as a dedicated CLI command for Codex, simulating the core functionality of the Claude Code `/pua:on` command. It enables 'PUA always-on mode' by managing the `~/.pua/config.json` file. This utility is essential for maintaining continuous operational readiness and managing coding assistant feedback cycles.

This alias serves as a specialized Codex CLI interface for implementing the P7 protocol, mimicking the workflow of a P7 senior engineer. It enforces a highly structured development cycle: detailed planning, robust implementation, mandatory self-review (via three critical questions), and final, verifiable completion markers. Ideal for complex coding tasks requiring maximum rigor and accountability.

A specialized CLI alias designed to implement the PUA P9 protocol, simulating the role of a P9 tech lead for AI coding assistance (Codex/Claude Code). This tool helps users structure complex development tasks by focusing on writing precise prompts, managing execution steps, and strictly adhering to a defined professional protocol, minimizing direct code implementation unless reassigned. It ensures rigorous process management and evidence-based output.

This CLI alias acts as a dedicated command to disable the PUA always-on mode feature, typically used within AI coding environments like Claude Code. By modifying the user's configuration file (`~/.pua/config.json`), it sets `always_on=false` and `feedback_frequency=0`. This function provides users with granular control, effectively turning off continuous monitoring or background feedback mechanisms managed by the AI assistant.

Use this skill to deeply inspect and audit the entire process of local AI coding-agent sessions. It tracks critical operational metrics such as token consumption, cost spikes, tool failures, latency anomalies, and overall health scores. Ideal for debugging unreliable code runs, establishing CI/CD health gates, or comparing different attempts to ensure semantic accuracy.

TokenWise is an advanced skill for Claude Code that automatically routes coding subtasks to the most cost-effective model (Haiku, Sonnet, or Opus). It logs every task's token usage and real cost, allowing users to audit spending and perform A/B tests to validate model tier efficiency, ensuring cost savings without sacrificing quality.