Automatically activated when user asks how something works, wants to understand unfamiliar code, needs to explore a new codebase, or asks questions like "where is X implemented?", "how does Y work?", or "explain the Z component"
SkillScore65Maintained
Automatic
◆Coding pack
Last verified 9/3/2026
What it changes
Passed 12 of 19 test cases with the skill, and 11 of 19 without it.
Without skill57%
With skill65%
Checks it passed that the model alone missed
The response outlines documenting execution flow and data transformations
The response outlines starting at the entry point to inspect initialization and imports
The response outlines documenting findings by building a mental model and tracking relationships between components
2 fixed13 passed either way8 still failing0 broke
Pass-rate lift
+9 pp
Tokens
+110%
Agent turns
0%
Response time
-10%
Model gemini-3.6-flashMeasured 9/3/2026v115 of 23 cases passed with skill · 19 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 2 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 2
Prompt
After initial reconnaissance on a web repo, we are starting Phase 2 targeted investigation to find where authentication tokens are parsed. Instead of jumping straight into line-by-line debugging without a plan, what sequence of steps defines Phase 2 targeted investigation?
What was checked, without and with the skill
Checked
Without skill
With skill
The response outlines searching for relevant code using Grep for keywords or Glob for file patterns
Passed
Passed
The response outlines reading key files starting from entry points to track execution flow and data transformations
Passed
Passed
The response outlines documenting findings by building a mental model and tracking relationships between components
Missed
Passed
case-12 · gemini-3.6-flash · 9/3/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md403 lines9 filesincludes scripts
Investigating Codebases
You are an expert code investigator with deep experience exploring and understanding unfamiliar codebases. This skill provides systematic investigation techniques to quickly understand code structure, patterns, and implementation details.
Your Capabilities
Structural Analysis: Map directory structures and identify architectural patterns
Dependency Tracing: Follow imports, function calls, and data flows
Pattern Recognition: Identify naming conventions, design patterns, and coding styles
Entry Point Discovery: Find main entry points, initialization code, and key workflows
Documentation Mining: Locate and synthesize existing documentation, comments, and READMEs
When to Use This Skill
Claude should automatically invoke this skill when:
The user asks "how does [feature/component] work?"
Questions about code location: "where is [X] implemented?"
Requests to explain unfamiliar code or systems
Tasks requiring exploration of unknown codebases
Questions about code organization or structure
Tracing execution flows or data paths
Understanding integration points between components
Investigation Methodology
Phase 1: High-Level Reconnaissance
1. Identify project type and structure
- Check package.json, Cargo.toml, setup.py, etc.
- Review top-level README and documentation
- Note framework/language patterns
2. Map directory organization
- Identify src/, lib/, app/ patterns
- Locate test directories
- Find configuration files
- Note special directories (scripts, tools, etc.)
3. Discover entry points
- Main files (main.js, index.ts, __init__.py, etc.)