spire-studio

Search and install agent skills from ClawHub, the public skill registry.

Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.

Generate images and iteratively edit saved image artifacts.

Sustained objectives via long_task / complete_goal — idempotent goal wording, project-style modular work, early web/doc research, Runtime Context metadata.

Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.

Summarize or extract text/transcripts from URLs, podcasts, and local files (great fallback for “transcribe this YouTube/video”).

Remote-control tmux sessions for interactive CLIs by sending keystrokes and scraping pane output.

One-time setup wizard for the CloakBot upgrade skill. Triggers: setup update, configure update, 切设置更新, 初始化更新.

Use when adding or documenting a local supervised fine-tuning dataset for Figaro LLM PEFT runs. Explains how to place JSONL files under datasets/llm, supported prompt_completion and messages formats, frontend sft.dataset_path discovery, and verification.

Use when adding or documenting a local LLM model for Figaro. Explains how to place Hugging Face-compatible model directories under models/llm, how Figaro discovers them as llm.base_model options, and how to verify the frontend/runtime selection.

FedArena Arena — evaluate user-designed FL attacks/defenses against a standardized benchmark matrix. Supports both prompt mode (describe your idea → Claude implements → auto-evaluate) and file submission mode.

Run FL benchmark experiments from natural language. Parses user intent, plans experiments, runs them via the research runner, and reports comparison results.