SONG BONGJUNE
Community@80portisfound
현재 지식 모델링 공부중
Agent Skills by SONG BONGJUNE
Showing 148 vetted skills indexed across 1 GitHub repositories.
google_meet
Joins Google Meet calls with a Playwright bot to transcribe captions and speak via OpenAI Realtime.
parallel-cli
Runs Parallel web search, extraction, deep research, enrichment, FindAll, and monitoring via CLI.
gitnexus-explorer
Index a codebase with GitNexus and serve an interactive knowledge graph via web UI.
qmd
Search local markdown notes and documents using hybrid BM25, vector, and LLM reranking retrieval.
drug-discovery
Query ChEMBL, PubChem, OpenFDA, and OpenTargets APIs for compound screening and drug-likeness analysis.
duckduckgo-search
Search the web via DuckDuckGo for text, news, images, and videos without an API key.
scrapling
Scrapes web pages via HTTP, stealth browser automation, and spider crawling with Scrapling.
bioinformatics
Indexes and fetches 400+ bioinformatics skills from bioSkills and ClawBio repositories on demand.
domain-intel
Performs passive domain reconnaissance including subdomain discovery, SSL inspection, WHOIS lookups, and DNS queries.
docker-management
Manage Docker containers, images, volumes, networks, and Compose stacks via CLI commands.
inference-sh-cli
Runs 150+ cloud AI apps for image, video, audio, and search via the infsh CLI.
sherlock
Search usernames across 400+ social networks using the Sherlock OSINT CLI.
oss-forensics
Investigate GitHub repositories for supply chain compromise using multi-source forensic evidence collection.
1password
Configure 1Password CLI and read or inject secrets into commands.
fitness-nutrition
Search exercises and food nutrition data via wger and USDA APIs, and compute fitness metrics offline.
neuroskill-bci
Reads real-time EEG and biometric metrics from a NeuroSkill BCI session to inform responses.
nemo-curator
Curate LLM training datasets with GPU-accelerated deduplication, quality filtering, and PII redaction.
huggingface-accelerate
Adds distributed training support to PyTorch scripts with four lines of code.
chroma
Store embeddings and metadata in Chroma for vector search and RAG applications.
simpo-training
Train LLMs with reference-free SimPO preference optimization using HuggingFace alignment-handbook configs.
slime-rl-training
Guides LLM post-training with reinforcement learning using Megatron-LM and SGLang.
guidance
Generate constrained LLM outputs using regex, grammars, and token healing with Guidance.
clip
Generate image and text embeddings for zero-shot classification and cross-modal retrieval.
lambda-labs-gpu-cloud
Launch and manage Lambda Labs GPU cloud instances for ML training and inference.
Frequently Asked Questions About SONG BONGJUNE
FAQPage SchemaWhat tasks can I accomplish with 80portisfound's skill registry?▼
You can fine-tune and align LLMs (LoRA, DPO, GRPO, SimPO), serve models via vLLM or llama.cpp, build RAG systems on Chroma/Qdrant/Pinecone/FAISS, generate images, video, and music, scrape websites, run OSINT investigations, manage GitHub pull requests, and control apps like Spotify, Notion, and Google Workspace.
Who is the target audience for these skills?▼
ML engineers running distributed training and post-training, developers building MCP servers and coding delegations, security researchers doing OSINT and repo forensics, data scientists curating training corpora with NeMo Curator, and creators producing diagrams, comics, music, and video through generative pipelines.
How do I install and run these skills in practice?▼
Skills load inside the Hermes Agent runtime on macOS or Linux. Most declare pip dependencies (torch, transformers, vllm) or CLI prerequisites (curl, jq, gh, ddgs) checked at runtime; some require environment variables such as NOTION_API_KEY, SHOPIFY_ACCESS_TOKEN, or a 1Password service account token.
Are these skills open source and free to use?▼
Nearly all skills carry an MIT license and are free to use and modify. The PowerPoint skill is the exception, distributed under a proprietary license defined in its own LICENSE.txt. External services like Pinecone, Lambda Labs, or Twilio may incur their own usage costs.
What prerequisites and dependencies do the skills require?▼
Common requirements include Python 3 with PyTorch and Transformers for ML skills, NVIDIA GPUs with RAPIDS for NeMo Curator, Blender 4.3+ for blender-mcp, macOS for Apple integrations like iMessage and FindMy, and API tokens for services such as Linear, Airtable, Canvas, and USDA FoodData Central.