Use for file-heavy or codebase-heavy tasks where the user wants to save tokens or avoid reading too much: finding relevant files, understanding a repo, editing large projects, auditing configs, comparing generated artifacts, or working with large logs, docs, PDFs, spreadsheets, or source trees.
Use for file-heavy or codebase-heavy tasks where the user wants to save tokens or avoid reading too much: finding relevant files, understanding a repo, editing large projects, auditing configs, comparing generated artifacts, or working with large logs, docs, PDFs, spreadsheets, or source trees.
File Context Scout
Use this skill to find the smallest useful file context before editing or explaining.
Search Order
Inspect repo status and top-level layout.
Search names with rg --files or native directory listing.
Search content with rg or git grep.
Read manifests and entry points before deep modules.
Open only the line ranges or sections needed.
Expand outward from imports, tests, routes, or call sites.
File Handling Rules
Never read dependency folders, caches, logs, or generated files unless the task is about them.
Prefer git diff for changed files.
Prefer AST, language server, or semantic tools when available. For Claude, use Serena when installed for symbol navigation and references.
For huge logs, sample head/tail and search error signatures first.
For PDFs/docs/spreadsheets, use available document tools or parsers before dumping raw text.
Editing Pattern
Before edits, record:
files likely involved
why each file matters
exact behavior being changed
tests or validation commands to run
After edits, summarize only changed behavior and validation results. Avoid restating whole files.
