Hermes Agent 默认会随附位于 optional-skills/ 目录下的可选技能,但这些技能默认处于关闭状态。如需使用,需手动进行安装:
hermes skills install official/<category>/<skill>
例如:
需将整个输入内容完整翻译,不得提前终止。
hermes skills install official/blockchain/solana
hermes skills install official/mlops/flash-attention
以下列出的每项技能均对应一个专门页面,其中详细介绍了其定义、配置方法及使用方式。
如需卸载:
hermes skills uninstall <skill-name>
| 技能 | 描述 |
| antigravity-cli | 操作 Antigravity CLI (agy):插件管理、身份认证及沙箱环境功能。 |
| blackbox | 将编程任务委托给 Blackbox AI 多模型 CLI 处理。 |
| grok | 将编程任务交由 xAI Grok Build CLI 执行(支持功能开发与 Pull Request 提交)。 |
| honcho | 配置并排查 Hermes 所使用的 Honcho 内存相关问题。 |
| openhands | 将编程任务委托给 OpenHands CLI 处理(支持模型无关操作及 LiteLLM 功能)。 |
| 技能 | 描述 |
| evm | 仅读型 EVM 客户端:可查询 8 条区块链上的钱包信息、代币数据及 Gas 费用。 |
| hyperliquid | 提供 Hyperliquid 市场数据、账户历史记录及交易审核功能。 |
| solana | 以美元为单位查询 Solana 网络上的钱包、代币、交易记录及 NFT 信息。 |
| 技能 | 描述 |
| 1-3-1规则 | 1-3-1决策简报:明确问题、列出三种方案,并选定一个最终选项。 |
| Skill | Description |
| ascii-art | ASCII art: pyfiglet, cowsay, boxes, image-to-ascii. |
| audiocraft-audio-generation | AudioCraft: MusicGen text-to-music, AudioGen text-to-sound. |
| baoyu-article-illustrator | Article illustrations: type × style × palette consistency. |
| baoyu-comic | Knowledge comics (知识漫画): educational, biography, tutorial. |
| comfyui | Generate images, video, and audio via diffusion workflows. |
| concept-diagrams | Generate flat, minimal educational SVG visuals as HTML. |
| creative-ideation | Generate ideas via named methods from creative practice. |
| draw-your-font | Turn a handwriting photo into an installable TTF font. |
| excalidraw | Hand-drawn Excalidraw JSON diagrams (arch, flow, seq). |
| heartmula | HeartMuLa: Suno-like song generation from lyrics + tags. |
| hyperframes | Render MP4/WebM videos from HTML compositions. |
| impeccable | Frontend design guidance, upstream-maintained (impeccable). |
| kanban-video-orchestrator | Plan and run multi-agent video production pipelines. |
| meme-generation | Create meme PNGs from templates with Pillow text overlay. |
| pixel-art | Pixel art w/ era palettes (NES, Game Boy, PICO-8). |
| pretext | Build creative browser demos with DOM-free text layout. |
| simple-english | Rewrite text to ASD-STE100 Simplified Technical English. |
| sketch | Throwaway HTML mockups: 2-3 design variants to compare. |
| social-media-content-calendar | Plan multi-platform social campaigns: briefs to posting. |
| tldraw-offline | Drive and script tldraw offline canvases with an agent. |
| touchdesigner-mcp | Control TouchDesigner via twozero MCP. |
| unreal-mcp | Automate Unreal Engine editor scenes, actors, and renders. |
| 技能 | 描述 |
| agentmail | 当智能体需要使用 AgentMail CLI 邮箱收件箱时使用。 |
| 技能 | 描述 |
| 三表模型 | 在 Excel 中构建整合了资产负债表、利润表和现金流量表的财务工作表。 |
| 可比公司分析 | 在 Excel 中构建用于可比公司估值的工作表。 |
| DCF模型 | 在 Excel 中构建折现现金流估值工作表。 |
| Excel文档生成 | 使用 openpyxl 以无界面方式生成可审计的财务工作表。 |
| 杠杆收购模型 | 在 Excel 中构建包含内部收益率及投资回收期等指标的杠杆收购分析工作表。 |
| 并购模型 | 在 Excel 中构建用于分析并购带来的价值增值或稀释效应的工作表。 |
| Polymarket接口 | 查询 Polymarket 的市场信息、价格数据、订单簿及历史记录。 |
| PowerPoint文档生成 | 使用 python-pptx 以无界面方式生成 PowerPoint 演示文稿。 |
| 股票信息 | 通过 Yahoo 提供股票行情、历史数据、搜索功能、对比分析以及加密货币相关信息。 |
| Skill | Description |
| accelerate | Run PyTorch training across GPUs with minimal changes. |
| axolotl | Axolotl: YAML LLM fine-tuning (LoRA, DPO, GRPO). |
| chroma | Embedding database for RAG and semantic search. |
| clip | Zero-shot image classification and image-text search. |
| dspy | DSPy: declarative LM programs, auto-optimize prompts, RAG. |
| evaluating-llms-harness | lm-eval-harness: benchmark LLMs (MMLU, GSM8K, etc.). |
| faiss | Fast vector similarity search at billion scale. |
| flash-attention | Speed up long-sequence transformer training and inference. |
| guidance | Constrain LLM output with grammars; guarantee valid JSON. |
| huggingface-hub | HuggingFace hf CLI: search/download/upload models, datasets. |
| huggingface-tokenizers | Fast BPE/WordPiece tokenization and custom vocab training. |
| instructor | Structured LLM outputs validated with Pydantic. |
| lambda-labs | On-demand GPU cloud instances for ML training. |
| llama-cpp | llama.cpp local GGUF inference + HF Hub model discovery. |
| llava | Vision-language chat: VQA, captioning, image dialogue. |
| modal | Serverless GPU cloud for ML jobs and model APIs. |
| nemo-curator | Curate LLM training data: dedupe, filter, PII redaction. |
| obliteratus | OBLITERATUS: abliterate LLM refusals (diff-in-means). |
| outlines | Outlines: structured JSON/regex/Pydantic LLM generation. |
| peft | Fine-tune large LLMs with LoRA on limited GPU memory. |
| pinecone | Managed vector DB for production RAG and search. |
| pytorch-fsdp | Fully sharded data-parallel training for large models. |
| pytorch-lightning | Clean training loops with built-in distributed support. |
| qdrant | Vector search engine for production RAG systems. |
| saelens | Train sparse autoencoders to interpret model features. |
| segment-anything-model | SAM: zero-shot image segmentation via points, boxes, masks. |
| serving-llms-vllm | vLLM: high-throughput LLM serving, OpenAI API, quantization. |
| simpo | Reference-free preference alignment, simpler than DPO. |
| slime | RL post-training for LLMs with Megatron and SGLang. |
| stable-diffusion | Text-to-image generation, inpainting, and img2img. |
| tensorrt-llm | High-throughput LLM inference on NVIDIA GPUs. |
| torchtitan | Pretrain LLMs at scale with PyTorch 4D parallelism. |
| trl-fine-tuning | TRL: SFT, DPO, GRPO, RLOO reward modeling for LLM RLHF. |
| unsloth | Unsloth: 2-5x faster LoRA/QLoRA fine-tuning, less VRAM. |
| weights-and-biases | W&B: log ML experiments, sweeps, model registry, dashboards. |
| whisper | Transcribe and translate speech in 99 languages. |
| Skill | Description |
| bioinformatics | Gateway to 400+ genomics and computational biology skills. |
| blogwatcher | Monitor blogs and RSS/Atom feeds via blogwatcher-cli tool. |
| darwinian-evolver | Evolve prompts/regex/SQL/code with Imbue’s evolution loop. |
| domain-intel | Passive recon of subdomains, SSL certs, WHOIS, and DNS. |
| drug-discovery | Drug discovery: ChEMBL search, drug-likeness, interactions. |
| duckduckgo-search | Free keyless web, news, and image search via ddgs. |
| gitnexus-explorer | Serve an interactive codebase knowledge graph web UI. |
| osint-investigation | Follow the money via public records and sanctions data. |
| parallel-cli | Agent-native web search, deep research, and enrichment. |
| pinecone-research | Agent RAG and long-term memory with Pinecone. |
| qmd | Hybrid local search over notes, docs, and transcripts. |
| research-paper-writing | Write ML papers for NeurIPS/ICML/ICLR: design→submit. |
| rss-feeds | Read RSS, Atom, JSON feeds; discover feeds behind a page. |
| scrapling | Scrape sites with stealth browsing and Cloudflare bypass. |
| searxng-search | Free keyless meta-search aggregating 70+ engines. |
| 技能 | 描述 |
| openhue | 通过 OpenHue CLI 控制 Philips Hue 灯具、场景及房间。 |
| 技能 | 描述 |
| yuanbao | Yuanbao(元宝)相关功能:@提及用户,查询信息或成员列表。 |
若要向该仓库添加新的可选技能:
- 在
optional-skills/<类别>/<技能名称>/ 下创建目录;
- 添加包含标准 frontmatter 信息的
SKILL.md 文件(包括名称、描述、版本及作者信息);
- 将相关辅助文件放入
references/、templates/ 或 scripts/ 子目录中;
- 提交 Pull Request——该技能将在合并后被收录至此目录,并拥有独立的文档页面。