Claude-skill-registry context-initializer

Automatically invokes init-explorer agent when project context is empty or unknown.

install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/context-initializer" ~/.claude/skills/majiayu000-claude-skill-registry-context-initializer && rm -rf "$T"
manifest: skills/data/context-initializer/SKILL.md
source content

Context Initializer

This skill detects when Claude Code lacks project context and automatically invokes the init-explorer agent to gather it.

When to Invoke This Skill

Invoke this skill when ANY of these conditions are true:

  1. No project context: You don't know what this project is about, its tech stack, or its purpose
  2. Missing session history:
    claude-progress.txt
    doesn't exist or hasn't been read
  3. Missing architect's digest:
    architects_digest.md
    doesn't exist or hasn't been read
  4. User asks context-dependent questions: The user asks about the project but you have no context
  5. Starting a new task: Beginning work on a feature without understanding the codebase

How to Check for Empty Context

Before invoking, verify context is actually missing:

# Check if session files exist
ls -la claude-progress.txt architects_digest.md 2>/dev/null || echo "Context files missing"

Invocation

When context is empty, invoke the init-explorer agent:

Task(subagent_type="init-explorer", prompt="
Gather project context for this codebase.

next_agent: none
task: Explore and document project structure, tech stack, and patterns

After exploration, return a summary of:
1. Project purpose and description
2. Tech stack (languages, frameworks, databases)
3. Key directories and their purpose
4. Testing setup
5. Build/run commands
")

What init-explorer Will Do

The init-explorer agent will:

  1. Orient: Run
    pwd
    ,
    ls -la
    ,
    git log
    ,
    git status
  2. Read History: Check
    claude-progress.txt
    for previous sessions
  3. Read Digest: Check
    architects_digest.md
    for task state
  4. Explore: Use Explore subagent for deep codebase analysis
  5. Create Files: Initialize
    architects_digest.md
    if missing
  6. Update Progress: Log the exploration session

Expected Output

After init-explorer completes, you will have:

  • Understanding of the project's tech stack and purpose
  • Knowledge of coding patterns and conventions
  • Awareness of test setup and build commands
  • Session logged in
    claude-progress.txt
  • Task tracking initialized in
    architects_digest.md

Example Usage

Scenario: User asks "How does authentication work in this project?"

Before: You have no context about the project.

Action: Invoke this skill to run init-explorer.

After: You understand the project uses Flask with JWT authentication, tests are in

tests/
, and the auth module is in
src/auth/
.

Do NOT Invoke When

  • You already have project context from earlier in the conversation
  • The user explicitly said not to explore
  • You're in the middle of a task that already has context
  • The init-explorer agent has already run in this session