hierarchical-context-docs
How to write and maintain architecture docs that present a system to an agent in progressive detail, instead of dumping the whole codebase.
Skill by utilitydelta
Chalk Labs
Need help deploying hierarchical-context-docs?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/utilitydelta-hierarchical-context-docs
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: writing, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/hierarchical-context-docs
What's inside
- Skill name
hierarchical-context-docs- Description
- How to write and maintain architecture docs that present a system to an agent in progressive detail, instead of dumping the whole codebase. Agents implementing new code features or refactoring work may need to update these documents.
- License
- MIT
- Instructions
- 24 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Hierarchical Context Docs Some code bases are too complex to search directly. Maintain a series of .md files in a `docs/` folder that has a hierarchical structure. These are not architectural design records. They represent the current state of the system. Everything is in source control, ruthlessly prune out-of-date documentation. ## The shape A tree of markdown files, each level more detailed than the last, each linking down to the next. - **L0 - root.** - One `docs/ARCHITECTURE.md` file. What the system is, what it is explicitly NOT, system-wide invariants, the map of subsystems, a link per subsystem. - **L1 - subsystem** - One file per major subsystem: its responsibility, its contract with the rest, its invariants, links to deeper design where it exists. - **L2 - subsystem details** - Only subsystems that need this depth get it. Do not include the code. The code speaks for itself. The code is the design. .md files only exist to compress this information. The agent follows the links it needs and stops when it has enough. It never holds the whole codebase in context to understand one corner. ## Rules - Markdown files link down the tree. They never link back up to parents. - Follow the `writing-style` skill. Preference `compressed-mode` if possible. You are writing for agents and a technical audience. - Don't link to specific code files or line numbers. These change all the time. - Place system invariants at the right level. Placing them too high in the document hierarchy creates noise. Placing them too low risks them not being discovered by agents. - Only current system state. Do not include implementation phases, design history or anything that's no longer current state for the system. Delete anything that references old system state.
From utilitydelta/build-method, MIT.
More from utilitydelta/build-method
- 0.563score
coding-rules
utilitydelta
Use when writing or reviewing code in any language.
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
explainer-video
utilitydelta
Use when building an animated explainer of a real system with manim (3b1b's manimgl, pinned - not manim CE), or when asked for a video, animation, or motion…
SkillBFreshness: B★ 2Claude Code +9 - 0.557score
writing-tests
utilitydelta
Use when writing tests, changing tests, or implementing code that needs tests.
SkillBFreshness: B★ 2Claude Code +9 - 0.557score
persona-driven-design
utilitydelta
Distill who the product is for before any design, architecture, or code.
SkillBFreshness: B★ 2Claude Code +9 - 0.554score
implementation-loop
utilitydelta
Use implementation loop to take a goal.md file and implement it in code.
SkillBFreshness: B★ 2Claude Code +9 - 0.551score
writing-style
utilitydelta
Writing style for docs, comments, commit messages, and PR text.
SkillBFreshness: B★ 2Claude Code +9 - 0.507score
performance-discipline
utilitydelta
Treat performance as a first-class invariant, measured empirically, not guessed.
SkillCFreshness: C★ 2Claude Code +9 - 0.504score
scout
utilitydelta
Use before handing a goal to implementation, or when a goal or UX journey is being authored or feels under-tested.
SkillCFreshness: C★ 2Claude Code +9 - 0.501score
compressed-mode
utilitydelta
Use this way in the user ask for a compressed response.
SkillCFreshness: C★ 2Claude Code +9
Signals
- Stars
- 2
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Jul 25, 2026 (3 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/utilitydelta/build-method
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from utilitydelta
- 0.563score
coding-rules
utilitydelta
Use when writing or reviewing code in any language.
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
explainer-video
utilitydelta
Use when building an animated explainer of a real system with manim (3b1b's manimgl, pinned - not manim CE), or when asked for a video, animation, or motion…
SkillBFreshness: B★ 2Claude Code +9 - 0.557score
writing-tests
utilitydelta
Use when writing tests, changing tests, or implementing code that needs tests.
SkillBFreshness: B★ 2Claude Code +9 - 0.557score
persona-driven-design
utilitydelta
Distill who the product is for before any design, architecture, or code.
SkillBFreshness: B★ 2Claude Code +9 - 0.554score
implementation-loop
utilitydelta
Use implementation loop to take a goal.md file and implement it in code.
SkillBFreshness: B★ 2Claude Code +9 - 0.551score
writing-style
utilitydelta
Writing style for docs, comments, commit messages, and PR text.
SkillBFreshness: B★ 2Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get hierarchical-context-docs running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
hierarchical-context-docs skill. Agent Mart. https://agentmart.dev/skills/utilitydelta-hierarchical-context-docs (last updated 2026-10-24)