Human Context Bundle
A machine-readable, vendor-neutral personal knowledge standard (OKF v0.2) designed to seamlessly ground AI coding assistants with your preferences, goals, and active projects.
Overview & Vision
AI coding assistants like Google Antigravity, Claude Code, and GitHub Copilot perform dramatically better when they understand who you are, how you think, and what you're currently working on.
The Human Context Bundle implements the Open Knowledge Format (OKF v0.2) standard to solve context fragmentation across different AI tools by providing a single, git-versioned knowledge repository.
Grounded Personalization
Tailors code style, tone, and explanations to your exact cognitive style and senior skillset.
Active Focus Alignment
Prioritizes your current goals, active sprint items, and strategic roadmap milestones.
Vendor-Neutral Standard
Compatible out-of-the-box with Google Antigravity, Anthropic Claude, and GitHub Copilot.
5-Step Quickstart
This repository will contain detailed personal information. We recommend keeping your repository Private or storing it locally on your computer.
1. Fork the Repository
Click the Fork button on GitHub to create your copy under your personal account.
2. Clone Locally
git clone https://github.com/catlikeflyer/human-context-bundle.git
cd human-context-bundle3. Fill In Your Context
Open the repository in your favorite editor and populate the markdown files in core/, state/, roadmaps/, and projects/. Whenever you add a new file, register it in the root index.md so AI agents can discover it.
4. Point Your AI Agents to the Bundle
Use pre-configured integrations in .agents/, .claude/, and .github/ to automatically ingest context into your AI coding tools.
5. Keep It Current
Commit updates regularly as your quarterly goals, active projects, or technical focus shift.
Bundle Index (index.md)
The root index.md is the master manifest of the entire bundle. It lists every context file across all directories in a single, structured table.
AI agents and tools are instructed to read index.md before navigating the rest of the bundle. This means any file not registered here may be invisible to your AI assistant. Always keep it up to date.
The index is organized by section — core/, state/, roadmaps/, and projects/ — with each row containing the file path, OKF type, and a short description. When you add a custom context file (e.g., core/hobbies.md), add a corresponding row to the appropriate section of index.md.
| File | Type | Description |
|------|------|-------------|
| `core/hobbies.md` | `core/hobbies` | Personal interests and creative outlets |Core Persona (core/)
The core/ directory houses your foundational background, skillset, education, and personality metrics. These files rarely change.
background.md
Core identity, career origins, and foundational principles.
experience.md
Professional career timeline, tech stack expertise, and domain competencies.
personality.md
Communication preferences, cognitive style, and collaboration norms.
education.md
Formal degrees, certifications, self-directed studies, and continuous learning.
All markdown files retain OKF v0.2 compliant YAML frontmatter header metadata at the top of the file.
The bundle is not limited to the default files. You are free to create any context files that make sense for you. For example, add a core/hobbies.md for personal interests, a core/values.md for guiding principles, or a state/health.md for wellness goals. Add valid OKF v0.2 frontmatter, then register the file in the root index.md so AI agents can discover it.
---
type: core
category: personality
title: Personality & Collaboration Profile
version: 0.2
---Dynamic State (state/)
The state/ directory contains dynamic, evolving priorities. AI agents consult these files to stay aligned with your immediate focus.
goals.md: Short-term and medium-term quarterly objectives.professional-focus.md: Current technical challenges and skill acquisition.personal-focus.md: Lifestyle priorities and personal development habits.future-plan.md: 3-5 year trajectory and long-term vision.
Strategic Learning Paths (roadmaps/)
Houses detailed execution paths for mastering new technologies or domain subjects.
Includes an index.md registry and a reusable template-roadmap.md to quickly launch new roadmap tracks.
Active & Portfolio Projects (projects/)
Tracks active side-projects, incubations, and archived repositories. AI agents inspect project files to understand architectural constraints, tech stacks, and current sprint backlogs.
AI Agent Calling & Grounding Guide
The repository includes native configuration adapters for major AI platforms:
1. Google Antigravity
Equipped with a custom agent skill in .agents/skills/context-filler/SKILL.md. When activated, Antigravity automatically reads core and state context before generating solutions.
2. Anthropic Claude
Configured with a custom subagent in .claude/agents/context-filler.md.
3. GitHub Copilot
Includes custom repository instructions in .github/copilot-instructions.md and an agent prompt in .github/agents/context-filler.agent.md.
CI/CD & Automated Validation
The repository includes pre-configured GitHub Actions workflows:
Frontmatter Linter (frontmatter-lint.yml)
Runs a lightweight Python validator (.github/scripts/lint_frontmatter.py) on every push and pull request to ensure all markdown files contain valid OKF v0.2 headers.
GitHub Pages Deploy (deploy-docs.yml)
Automatically packages and deploys this docs/ directory to GitHub Pages upon merging changes to main.
Privacy & Security Best Practices
Never commit secrets, passwords, private API keys, or confidential employer data to your context bundle.
Keep your bundle repository Private on GitHub, or use it exclusively on your local workstation.