Skill UI

Secure AI coding agents with permission boundaries, secret protection, code review gates, and safe sandbox configurations for team environments. Covers Claude Code, Cursor, Copilot, and Codex security implementations including CLAUDE.md configuration, command allowlists, filesystem access controls, and pre-commit hooks for secret scanning.

Orchestrates bounded coding tasks to Aider CLI, managing briefs, diffs, and reviews. Supports local models via OpenAI-compatible endpoints, ensuring offline operation and no auto-commits.

Converts requirements into precise, dependency-aware implementation plans. Creates detailed checklists, defines milestones, and maps critical paths for development teams. Ensures tasks are atomic and verifiable before coding begins, bridging the gap between requirements and execution.

Forensic analyzer for AI coding sessions. Identifies root causes of scope drift and rework by classifying user specs, agent errors, and repo friction. Generates evidence-backed reports to improve future prompts and project health.

This skill reviews repository architecture using Ontoly Software Graph and MCP capabilities. It helps explain architecture, module boundaries, package topology, service ownership, and architectural risks. It utilizes capabilities like ExplainArchitecture, GraphStatistics, FindCycles, and FindDependencies to provide evidence-based analysis. Suitable for coding agents requiring architectural insights without manual file searching.

AWT provides AI coding tools with visual perception and interaction capabilities for web applications. It executes declarative YAML test scenarios using Playwright, featuring visual matching via OpenCV and OCR without relying on stable DOM selectors. Supports platform auto-detection for frameworks like React and Vue, includes structured failure diagnosis, and maintains a learning database for failure patterns. Ideal for end-to-end testing workflows requiring visual assertions and AI-assisted debugging.

This skill reconstructs the detailed workflow behind a specific coding artifact (like a feature or PR). By analyzing local `ax` agent traces, it connects commits, sessions, skills, and tool calls into a concise, evidence-grounded narrative. It is used when needing to understand the step-by-step process and key decisions that led to a result.