What deep-init runs automatically, and when. A hook is a command Claude Code fires at a fixed moment, without you asking for it.
Claude Code
> /plugin marketplace add deepfusionlabs/deep-init> /plugin install deep-init@deepfusionlabs-deep-init
Cursor
Cursor also runs Claude Code hooks from .claude/settings.json, so copy these hooks there to use them in Cursor.
What fires, and when
SessionStart
Fires once when a session begins, and again after a context compaction. It is where a plugin sets up its environment, or restores state the compaction dropped.
"${CLAUDE_PLUGIN_ROOT}/skills/deep-init/assets/session-start.sh"UserPromptSubmit
Fires before Claude sees each prompt you send. A plugin can use it to inject context, so the same instruction reaches the model every turn instead of only at session start.
"${CLAUDE_PLUGIN_ROOT}/skills/deep-init/assets/session-start.sh"
Claude Code plugin that writes grounded, verified context for your coding agent — every claim checked against your code. Lean CLAUDE.md + deep docs, 100% local, MIT.
Stars on GitHub
Gives you a guide to setting up Claude Code, covering global instructions, MCP servers, custom commands, hooks and skills, plus tools for generating commit messages and auditing code for security issues. This version is marked obsolete, with a newer starter kit recommended instead.
@thedecipheristGives your coding agent skills to author, maintain, validate and visualize Open Knowledge Format bundles, portable markdown knowledge that both your team and your agents can read. Includes a backfill skill for building a bundle from an existing repository's git history and session transcripts.
@scaccogattoAutomatically creates, enhances and maintains CLAUDE.md files for Claude Code projects, so you don't have to write and update them by hand.
@alirezarezvaniTurns design docs and plan drafts into HTML pages you can comment on inline. Fires when you ask to review a design, preview a plan, or render visual output.
@u-ichiA Claude Code plugin that maintains FILETREE.md, a one-line description per file with content hashes for staleness detection. Lets the LLM grasp repo layout in a few hundred tokens before touching code.
@nekocode
@claudateKeeps you fluent in code that AI writes: each feature is designed, built with the reasoning behind each decision explained, and recorded as docs, decisions and review summaries, so the understanding stays with the code.
@baokhang83A Claude Code plugin that reads your codebase and draws its infrastructure and architecture with D2, then keeps those diagrams in step with the code as it changes. Command: /d2:diagram.
@heathdutton