How do I install the Intellidiff skill?

Run npx skillmds@latest add batteryshark/intellidiff in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Intellidiff safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Intellidiff?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Intellidiff free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Intellidiff?

batteryshark (@batteryshark) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Intellidiff by batteryshark — Agent Skill

Name: Intellidiff
Author: batteryshark

Intellidiff

Compare files or directory trees, detect exact duplicate files, calculate SHA-256 checksums, normalize text for semantic comparison, and read targeted line ranges. Use when determining whether artifacts are identical, finding orphaned or changed files between folders, deduplicating collections, inspecting backups, or explaining content differences without modifying the inputs.

batteryshark c069487 5 files · 15.4 KB Updated Sep 17, 2026

File contents

IntelliDiff

Use deterministic, read-only comparisons before making deduplication or merge decisions.

Commands

Run the portable entry point:

bin/intellidiff hash FILE
bin/intellidiff file LEFT RIGHT
bin/intellidiff file LEFT RIGHT --smart --ignore-whitespace --ignore-blank
bin/intellidiff folders LEFT RIGHT
bin/intellidiff duplicates FOLDER
bin/intellidiff lines FILE --start 20 --end 40 --context 3

Read references/reference.md for normalization semantics, hidden-file behavior, JSON output, and exit codes.

Workflow

Choose exact file comparison for byte identity and smart comparison only when specific textual differences should be ignored.
Compare directory trees before deciding which side is canonical. Review same-path changes separately from left-only and right-only files.
Use duplicate detection to find exact content groups. Hash equality is strong evidence for identical content, not authorization to delete a copy.
Inspect metadata, neighboring files, references, and provenance before moving or removing duplicates.
Report the comparison mode and every normalization applied so the meaning of “same” is explicit.

Rules

Never modify, move, or delete inputs.
Use SHA-256 rather than timestamps, names, or CRC checksums for identity grouping.
Do not call normalized text byte-identical.
Hidden files are skipped by default except when --include-hidden is explicit; .git and common dependency/build caches remain excluded.
Symlinks are not followed.
Treat unreadable files and decode failures as explicit errors, not silent differences.

Use agents/analyst.md when the comparison needs a human-readable merge or deduplication recommendation.

---name: intellidiffdescription: Compare files or directory trees, detect exact duplicate files, calculate SHA-256 checksums, normalize text for semantic comparison, and read targeted line ranges. Use when determining whether artifacts are identical, finding orphaned or changed files between folders, deduplicating collections, inspecting backups, or explaining content differences without modifying the inputs.---# IntelliDiffUse deterministic, read-only comparisons before making deduplication or merge decisions.## CommandsRun the portable entry point:```shbin/intellidiff hash FILEbin/intellidiff file LEFT RIGHTbin/intellidiff file LEFT RIGHT --smart --ignore-whitespace --ignore-blankbin/intellidiff folders LEFT RIGHTbin/intellidiff duplicates FOLDERbin/intellidiff lines FILE --start 20 --end 40 --context 3```Read [`references/reference.md`](references/reference.md) for normalization semantics, hidden-file behavior, JSON output, and exit codes.## Workflow1. Choose exact file comparison for byte identity and smart comparison only when specific textual differences should be ignored.2. Compare directory trees before deciding which side is canonical. Review same-path changes separately from left-only and right-only files.3. Use duplicate detection to find exact content groups. Hash equality is strong evidence for identical content, not authorization to delete a copy.4. Inspect metadata, neighboring files, references, and provenance before moving or removing duplicates.5. Report the comparison mode and every normalization applied so the meaning of “same” is explicit.## Rules- Never modify, move, or delete inputs.- Use SHA-256 rather than timestamps, names, or CRC checksums for identity grouping.- Do not call normalized text byte-identical.- Hidden files are skipped by default except when `--include-hidden` is explicit; `.git` and common dependency/build caches remain excluded.- Symlinks are not followed.- Treat unreadable files and decode failures as explicit errors, not silent differences.Use [`agents/analyst.md`](agents/analyst.md) when the comparison needs a human-readable merge or deduplication recommendation.

Frequently asked questions