# Basement Hacker

> Goes underground to reverse-engineer internals, read the source nobody reads, and explain how something actually works; invoke via /basement-hacker when documentation is missing or lying and you need ground truth from the implementation itself.

- Skill: `jakeschincariol/basement-hacker` (Agent Skill)
- Install (CLI): `npx skillmds@latest add jakeschincariol/basement-hacker`
- Raw SKILL.md: https://api.skillmd.com/api/skills/jakeschincariol/basement-hacker/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing
- Author: Jakeschincariol (https://skillmd.com/u/jakeschincariol)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/jakeschincariol/basement-hacker

---


# Basement Hacker

Down past the docs, past the abstractions, into the source nobody opens. This mode reverse-engineers how a thing actually works — not how the README claims it does — and comes back up with the real mental model.

## When to use
- Docs are missing, outdated, or contradicting the observed behavior.
- You need ground-truth understanding of a library, API, or system's internals.

## Behavior
1. Go to the source — read the actual implementation, not just the public API surface or docs.
2. Trace the real execution path: entry points, internal calls, state, and side effects.
3. Reconstruct the underlying data model and control flow from the code itself.
4. Test hypotheses against behavior — small experiments to confirm how it really responds.
5. Note where the documented behavior and the actual behavior diverge.
6. Surface the hidden gotchas: undocumented assumptions, magic constants, order dependencies.
7. Explain the mechanism in plain language with citations to the specific source you read.

## Output
A ground-truth explanation of how the thing actually works, citing the specific source, plus a list of doc-vs-reality discrepancies and hidden gotchas.