Hypergraph-Coding-Agent-Framework: Skill for Claude Code

Hyper-Document

This skill updates the full documentation suite after changes to the codebase, skills, or framework structure. It reads actual implementation before writing anything — existing documentation is treated as potentially stale.

---

When to use this skill

• After completing a feature, bug fix, refactor, or skill change.
• When README.md, CHANGELOG.md, docs/, or other docs are out of sync with the codebase.
• When the user explicitly runs /hyper-document.

---

Step 1 — Identify What Changed

Run the following to understand scope:

git log --oneline -10
git diff HEAD~1 --name-only

Identify:

• Which files were modified, added, or deleted.
• Whether the change is user-facing (affects README, CHANGELOG, docs) or internal (affects memory files, skills-info).
• Whether any skills were added, renamed, or removed.

---

Step 2 — Read Before Writing

CRITICAL: Do NOT trust existing documentation. Read the actual implementation.

• Read every changed file before updating any doc.
• Note discrepancies between existing docs and actual behavior.
• Never assume — verify.

---

Step 3 — Update Documentation by Target

Work through each target below. Skip targets that are clearly unaffected by the change.

---

README.md

Purpose: Entry point for new users and project overview.

Update when: Architecture changes, new features, new IDE support, install/uninstall changes, workflow changes, new slash commands, or directory structure changes.

Style rules:

• Lead with what it does, not how it works.
• Use H2 (##) for top-level sections, H3 (###) for subsections.
• Prefer tables for structured comparisons (IDE support, installed paths, commands).
• Use fenced code blocks with language tags for all commands and examples.
• Bullet lists for feature enumerations; numbered lists for sequential steps only.
• No marketing language. No "powerful", "robust", "seamless".
• Every slash command mentioned must use the hyper- prefix (e.g. /hyper-architect).
• Callout blocks (> **Note:**) for caveats, naming conventions, and non-obvious behaviors.