| name | trace-code |
| description | Deep code flow tracing with file:line references. Use when the user says "trace this", "how does this feature work end-to-end", "map the code flow", "what files are involved in", "walk through the execution". |
Trace Code
Trace a feature's implementation from entry point to output, through all layers.
Announce at start: "Tracing [feature/flow name]."
Process
- Find entry points — CLI commands, API endpoints, script main blocks
- Follow call chains — trace each function call with file:line references
- Map data flow — what goes in, how it transforms, what comes out
- Identify dependencies — external libraries, AWS services, config files
- Note cross-cutting concerns — error handling, logging, auth, retries
Output Format
Entry Point
file.py:42 — description of where execution starts
Execution Flow
file.py:42 → module.py:15 — what happens at each step
module.py:15 → aws.py:88 — data transformation or side effect
Key Components
component.py — responsibility summary
Dependencies
- External: boto3, rich, etc.
- Config: files or env vars read
Essential Files
List of files someone must read to understand this feature.
Guidelines
- Always include file:line references — never describe without pointing to code
- Trace the happy path first, then note error/edge case branches
- Keep it factual — describe what the code does, not what it should do