Babysitter codebase-research

Systematic codebase exploration following the Iron Law - understand the problem before exploring code. Four phases with file-finder and web-researcher agents.

install

source · Clone the upstream repo

git clone https://github.com/a5c-ai/babysitter

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/a5c-ai/babysitter "$T" && mkdir -p ~/.claude/skills && cp -r "$T/library/methodologies/rpikit/skills/codebase-research" ~/.claude/skills/a5c-ai-babysitter-codebase-research && rm -rf "$T"

manifest: library/methodologies/rpikit/skills/codebase-research/SKILL.md

Codebase Research

Overview

Perform systematic codebase exploration to understand how existing systems work. Follows the Iron Law: "Do NOT explore the codebase until the problem is understood."

When to Use

Implementation direction is clear but codebase understanding is needed
Investigating how an existing feature works before modifying it
Understanding dependencies and data flows before planning
Gathering context for a known goal

Process

Understand the request - Ask clarifying questions one at a time (purpose, specifics, scope, constraints, context). Do NOT read any files until confirmed.
Explore the codebase - Use file-finder agent, read in order, trace data flows, identify constraints.
Document findings - Write structured research document to
```
docs/plans/YYYY-MM-DD-<topic>-research.md
```
.
Transition - Ask: plan, continue research, or conclude.

Key Rules

Quotations from source material capped at 125 characters maximum
Only proceed to exploration after human confirms understanding
Use file-finder agent for initial file discovery
Use web-researcher agent for external context needs

Agents Used

```
agents/file-finder/
```
- Locates relevant files with suggested reading order
```
agents/web-researcher/
```
- Gathers external context when needed

Tool Use

Invoke via babysitter process:

methodologies/rpikit/rpikit-research