ddalcu

CEO skill for Claude Code — manages organizations via Agent Orcha REST APIs

Documentation for creating and modifying ORCHA resources (agents, workflows, skills, functions, knowledge stores)

CEO skill for agent-based organization leadership — ticket triage, delegation, resource creation, and work review

Browser automation — observe-act loop with refs

PII and OWASP security filtering rules

Execute JavaScript, shell commands, fetch web content, and control a browser

A skill that requires sandbox

A test skill for unit testing

Hook an app, game or script up to the local mlx-serve server for LLM chat, embeddings, image, speech, music, sound effect, video and 3D generation, and Laya/Kev/Clef typed decisions. Use when code should call mlx-serve.

mlx-serve pre-release validation checklist, CalVer versioning, release steps, and CHANGELOG style. Use when preparing or cutting a release, running pre-release validation, or writing CHANGELOG entries.

Turn llmprobe reports into the charts a perf PR embeds (one line per arm across context size, every point labelled, machine and model in the caption) and host them so the PR body renders them. Use when a PR makes a speed claim, or when asked for a chart of a…

mlx-serve benchmarking methodology — bench.sh/llmprobe usage, comparison-trap rules (same-methodology cells only, spec-decode variance, thermal lies, engine naming), perf-claim etiquette. Use before running benchmarks or making any performance claim.