Code Archaeologist — Bundle

The problem

Inheriting an undocumented legacy codebase or a repo from an acquired startup often leads to weeks of confusion. Jumping straight into refactoring without understanding the hidden dependencies and historical intent results in breaking changes and technical debt.

What it does

• Surveys unfamiliar codebases to map structure and dependencies before suggesting modifications.
• Cites specific file paths and line numbers for every claim to ensure evidence-based analysis.
• Categorizes findings into confidence tiers: confirmed facts, probable theories, and speculative inferences.
• Investigates version control logs and blame data to reconstruct the original intent behind complex logic.
• Recommends the smallest safe next step for onboarding rather than high-risk, large-scale rewrites.

Why this beats prompting it yourself

Generic LLMs often hallucinate structural details or suggest aggressive, destructive refactors of code they don't fully grasp. This skill enforces a read-only, evidence-first methodology that prevents premature code changes and ensures every observation is grounded in the actual source files.

Use cases

• Onboarding to a legacy monolith where the original authors are no longer available.
• Audit acquired repositories to understand architectural patterns and hidden risks.
• Mapping dependencies before starting a migration or major feature update.
• Documenting tribal knowledge by analyzing commit history and code patterns.

Known limitations

Does not provide legal or license interpretations. Refuses to perform deletions or refactors during the initial survey phase without explicit confirmation.