Teams need reusable instructions that AI agents can find and follow quickly. This skill guides concise skill structure, discovery wording, examples, checks, and iterative improvement.

Claude rule files can fail when teams reuse Cursor fields or invalid glob syntax. This skill gives clear patterns for valid .claude/rules/ files, scoped paths, and migration cleanup.

Claude Code agents can fail when their frontmatter, persona, tools, or model settings are unclear. This skill gives structured guidance for creating focused agents that validate and work predictably.

Video editing prompts often fail when routed to the wrong model. This skill selects a RunComfy video model and applies targeted prompting guidance for the requested edit.

Image editing workflows often lose identity, framing, or batch consistency. This skill guides Claude, Codex, and Claude Code to use Nano Banana Edit through RunComfy with structured prompts and clear routing rules.

Teams need fast image drafts that follow layout, text, and aspect-ratio requirements. This skill guides Claude, Codex, and Claude Code to use Nano Banana 2 through RunComfy with practical prompt patterns.

Lip-sync workflows require choosing the right model for each media input. This skill routes video, portrait, audio, and script requests to the matching RunComfy endpoint.

Cropped images often need more space for banners, product pages, and social formats. This skill routes outpainting requests to RunComfy edit models and helps preserve the original subject.

Teams need quick custom music without managing a full audio production workflow. This skill guides Claude, Codex, or Claude Code through structured ElevenLabs Music prompts using the RunComfy CLI.

Pose and motion references are difficult to route to the right generation model. This skill selects RunComfy routes for pose transfer, ControlNet image generation, and related workflows.

Agent work often lives in local folders that teammates cannot inspect while it changes. This skill creates an Agentspace link for a chosen path so others can view, comment on, or edit the workspace.

Security teams need repeatable code review without building rules from scratch. This skill guides Semgrep scans, OWASP and CWE mapping, remediation, and CI integration.

Security teams need repeatable threat models that stay aligned with changing architecture. This skill helps Claude, Codex, and Claude Code guide pytm-based STRIDE analysis and report generation.

Security teams need portable detection logic that works across SIEM platforms. This skill helps create Sigma rules, map them to ATT&CK, and prepare them for hunting or compliance workflows.

Container images often ship with vulnerable packages that are hard to prioritize. This skill helps Claude, Codex, and Claude Code run Grype scans, interpret results, and add CI/CD gates.

API specifications can miss security requirements before implementation starts. This skill guides Claude, Codex, and Claude Code through Spectral linting, OWASP API checks, and governance workflows.

Teams often lose consistency when UI decisions, accessibility checks, and component rules are handled informally. This skill gives Claude, Codex, and Claude Code a concise operating guide for structured UI/UX review and documentation.

Inconsistent test planning and reporting can hide quality gaps. This skill gives Claude, Codex, and Claude Code a structured process for test design, mocking, execution checks, and reporting.

Teams using AI agents can lose context when tasks split across many workers. This skill gives Claude, Codex, and Claude Code a clear operating model for delegation, reporting, review, and safe closure.

Schema changes can break validators, downstream types, and migration paths. This skill guides Claude, Codex, and Claude Code through scoped AFI signal schema updates with tests and governance checks.