Skill UI

Local reference for applying Anthropic's official brand colors, Poppins/Lora typography, and accent styling to requested artifacts such as slides or documents, with fallback fonts and export verification.

This skill prevents AI coding agents from entering ineffective repair loops by establishing a bounded repair contract. It uses stable failure fingerprints, a three-attempt budget, and real-path verification to ensure progress. When fixes do not change observable states, it mandates shifting root-cause strategies rather than retrying. Includes rollback controls and negative testing to validate actual system behavior over agent claims.

BrowserAct is a browser automation CLI for AI agents, enabling real browser interaction, JavaScript-rendered extraction, screenshots, network capture, parallel sessions, verification handling, and human handoff. Ideal for tasks requiring authenticated state, dynamic content, or multi-session isolation.

This skill drives an existing browser instance using the Chrome DevTools Protocol (CDP). It is essential for complex web automation tasks that require visible interaction, JavaScript execution, or handling authenticated, stateful sessions. Use it when static data fetching is insufficient, such as simulating user logins, interacting with iframes, or performing visual verification. Supports headless and remote environments.

Learn how to use a Win32 build of BusyBox to run many standard UNIX command line tools on Windows without WSL. Includes architecture detection, download instructions, and security verification.

Clarity Gate is a pre-ingestion verification system designed for Retrieval-Augmented Generation (RAG) workflows. It ensures documents entering knowledge bases are properly qualified by checking for uncertainty markers and distinguishing assumptions from facts. This skill produces Clarity-Gated Documents (CGD) and validates Source of Truth (SOT) files. It helps prevent hallucinations by enforcing epistemic quality standards before data ingestion. Note that it verifies document form rather than factual truth, requiring human-in-the-loop validation for accuracy.

Drives a coding task end-to-end against explicit acceptance criteria: minimal implementation, local tests, PR review loop, dev deploy and runtime verification, then evidence-based completion reporting.

Perform authorized static analysis on source code to identify security vulnerabilities using tools like Semgrep and CodeQL. This workflow covers pattern hunting, dangerous API detection, and verification of vulnerability fixes. Essential for white-box audits and PR security reviews.

This skill enables orchestrating agents to delegate bounded coding tasks to the OpenAI Codex CLI. The orchestrator drafts the brief, runs the command, and maintains full responsibility for reviewing diffs and committing changes. It prioritizes safety by preventing automatic commits, requiring explicit user requests and manual gate verification.

This skill applies a rigorous, evidence-first discipline to coding tasks, inspired by Fable-like best practices. It guides agents through a structured workflow: inspecting the workspace, tracking goals and findings, grounding all conclusions in verifiable evidence, and executing thorough verification gates before declaring completion. Ideal for complex development, debugging, or adapting cross-model prompt guidance into robust, verifiable project rules.

Execute discrete, bounded coding, review, or verification tasks using the Codex CLI as an isolated subagent. This tool is ideal for tasks requiring explicit file ownership, such as isolated refactoring, feature implementation, or independent code verification, without relying on continuous conversational context. It runs in a safe sandbox environment.

Create new development tracks (features, bugs, chores, refactors) with detailed specifications and phased implementation plans. Helps organize work with acceptance criteria, dependencies, and verification steps for structured development workflows.