Columbia-1 Skill Library
A concise inventory of the capabilities currently available to Hermes in Davis's active profile.
★
Bottom line
This profile currently has 70 skills spanning research, productivity, software development, integrations, creative work, ML/AI tooling, and agent operations — including the new Columbia Pages publishing skill used to make this report.
Installed Skills
70
available now
Categories
15
plus general QA
Newest Addition
Pages
shareable reports
Productivity
| Skill | What it does |
|---|---|
| columbia-pages New | Publishes polished house-themed HTML reports and returns shareable links. |
| discord-response-style | Formats concise Discord/gateway replies with ping and truncation handling. |
| executor-google-resources | Reads approved Gmail, Calendar, and Drive sources through Executor. |
| executor-x-resources | Reads X/Twitter posts, timelines, accounts, and search through Executor. |
| maps | Handles geocoding, POIs, routes, and timezones. |
| nano-pdf | Edits PDF text, typos, and titles via the nano-pdf CLI. |
| notion | Retrieves and summarizes 10XN or Theo Notion content through Executor. |
| ocr-and-documents | Extracts text from PDFs and scanned documents. |
| slack-read-access | Sets up safe, mostly read-only Slack access across workspaces. |
| social-stats-digests | Creates recurring social performance snapshots, deltas, and highlights. |
| todoist | Inspects and manages Davis's Todoist tasks through Executor. |
Everyday workflows, connected resources, and shareable output.
Software Development
| Skill | What it does |
|---|---|
| debugging-hermes-tui-commands | Debugs Hermes TUI slash commands across Python, gateway, and Ink UI. |
| hermes-agent-skill-authoring | Authors and validates Hermes SKILL.md files. |
| hermes-s6-container-supervision | Modifies or diagnoses s6-overlay services in Hermes Docker images. |
| local-status-dashboards | Builds small local dashboards for developer machines and agent runtimes. |
| node-inspect-debugger | Debugs Node.js through the inspector and Chrome DevTools Protocol. |
| plan | Produces implementation plans without executing changes. |
| python-debugpy | Debugs Python with pdb and debugpy/DAP workflows. |
| requesting-code-review | Runs pre-commit quality, security, and review checks. |
| spike | Uses disposable experiments to validate ideas before building. |
| subagent-driven-development | Executes plans via delegated subagents with review stages. |
| systematic-debugging | Applies root-cause debugging before attempting fixes. |
| test-driven-development | Enforces red-green-refactor development discipline. |
| writing-plans | Writes bite-sized, path-specific implementation plans. |
Agent Operations & DevOps
| Area | Skill | What it does |
|---|---|---|
| Agents | codex | Delegates coding features and PR work to OpenAI Codex CLI. |
| Agents | hermes-agent | Configures, extends, and contributes to Hermes Agent itself. |
| Agents | hermes-profile-curation | Audits and simplifies profile skills, job dependencies, and routing. |
| Agents | hermes-usage-analytics | Answers grounded Hermes token, usage, and cost questions. |
| Agents | kanban-codex-lane | Runs Codex as an isolated implementation lane under Kanban control. |
| DevOps | kanban-orchestrator | Coordinates multi-worker Kanban decomposition and routing. |
| DevOps | kanban-worker | Guides Hermes workers through Kanban edge cases and pitfalls. |
| DevOps | project-secrets-management | Evaluates env-var and secret management for worktrees, agents, CI, and deploys. |
| DevOps | webhook-subscriptions | Sets up webhook-driven agent runs. |
Research, Knowledge & Connected Content
| Area | Skill | What it does |
|---|---|---|
| Research | arxiv | Searches arXiv papers by keyword, author, category, or ID. |
| Research | blogwatcher | Monitors RSS/Atom feeds and blogs. |
| Research | llm-wiki | Builds and queries Karpathy's interlinked LLM Wiki knowledge base. |
| Research | research-paper-writing | Writes ML papers targeting NeurIPS, ICML, and ICLR workflows. |
| Research | web-research | Researches the web with extraction, validation, and fallback disclosure. |
| Knowledge | gbrain-knowledge-base | Audits, maintains, upgrades, and queries Davis's local GBrain. |
| Media | youtube-content | Works with YouTube transcripts, summaries, transformations, and archives. |
| agentmail-inboxes | Operates agent-owned inboxes with safe access and inbound processing. | |
| Import | explore_data | Analyzes the configured ytvids PostgreSQL database. |
| Import | ytvids_sync | Syncs configured YouTube channels into the ytvids database. |
GitHub & Integrations
| Area | Skill | What it does |
|---|---|---|
| GitHub | codebase-inspection | Inspects codebase language and line-count composition. |
| GitHub | github-auth | Sets up GitHub HTTPS tokens, SSH keys, and gh login. |
| GitHub | github-code-review | Reviews PR diffs and submits inline comments. |
| GitHub | github-issues | Creates, triages, labels, and assigns GitHub issues. |
| GitHub | github-pr-workflow | Handles branches, commits, PRs, CI, and merges. |
| GitHub | github-repo-management | Clones, creates, forks, and manages repositories and releases. |
| MCP | native-mcp | Connects and registers MCP servers and tools. |
Creative & Design
| Skill | What it does |
|---|---|
| architecture-diagram | Creates dark-themed SVG architecture and infrastructure diagrams. |
| ascii-art | Makes ASCII art with figlets, cowsay, boxes, and image conversion. |
| ascii-video | Converts media to colored ASCII MP4 or GIF output. |
| claude-design | Designs one-off HTML artifacts such as landing pages or decks. |
| excalidraw | Creates hand-drawn Excalidraw architecture, flow, and sequence diagrams. |
| humanizer | Removes AI-isms and gives text a more natural voice. |
| popular-web-designs | Builds HTML/CSS using real product design-system references. |
| sketch | Produces quick HTML mockup variants for comparison. |
ML, Data & Model Tooling
| Area | Skill | What it does |
|---|---|---|
| Data | jupyter-live-kernel | Runs iterative Python exploration through a live Jupyter kernel. |
| ML | audiocraft-audio-generation | Generates music and audio with AudioCraft models. |
| ML | dspy | Builds and optimizes declarative LM programs and RAG systems. |
| ML | evaluating-llms-harness | Benchmarks LLMs with lm-eval-harness. |
| ML | huggingface-hub | Searches, downloads, and uploads models or datasets through Hugging Face. |
| ML | llama-cpp | Runs local GGUF inference and finds models on Hugging Face. |
| ML | obliteratus | Supports OBLITERATUS model refusal-ablation workflows. |
| ML | segment-anything-model | Performs zero-shot image segmentation using SAM. |
| ML | serving-llms-vllm | Serves LLMs at high throughput through vLLM. |
| ML | weights-and-biases | Tracks experiments, sweeps, and model registry workflows in W&B. |
Desktop & Quality Assurance
| Skill | What it does |
|---|---|
| macos-computer-use | Drives macOS desktop interactions in the background without stealing input focus. |
| dogfood | Performs exploratory web-app QA and produces evidence-backed bug reports. |