Diff

Go beyond raw git diff — follow dependency chains, categorize changes, and calculate blast radius.

Steps

1. Identify branches

Parse <branch1> and <branch2> (default: current vs main). Find common ancestor via git merge-base. Read .ccs/branches/<branch>.md if they exist.

2. Gather raw diff

git diff --stat <branch1>...<branch2> → file-level changes
git diff --shortstat <branch1>...<branch2> → totals
git diff --name-only <branch1>...<branch2> → clean file list
git log --oneline <branch1>...<branch2> → commit list

3. Analyze impact

Read .ccs/file-index.md → rank changed files
Read .ccs/project-map.md → find downstream dependents (1 level)
Flag entry points, configs, shared utilities
Categorize: feature (new files/exports), fix (small changes), refactor (renames/moves), config

4. Calculate blast radius

Direct changes: files in diff
First-order dependents: files importing changed files
Risk level: based on file ranks (S-rank = high risk)
Breaking change potential: exports removed, interfaces changed, types modified

5. Generate report

Output: change type, risk level, direct changes count, blast radius count. Table of changed files by importance (file, rank, +/-, dependents). Dependency impact table. Change categories breakdown. Commits list. Risk assessment (breaking changes, S-rank mods, shared utils, test coverage).

Update .ccs/branches/<branch-name>.md with analysis.

6. Log to `.ccs/task.md`

Append using template at .claude/skills/_ccs/templates/task-template.md: branches compared, direct changes, blast radius, risk level, change type.

Rules

Use --name-only and --stat — never full diff output
Read file-index and project-map once, not per file
Grep for import statements rather than reading entire files
Cap dependency walking at 1 level
For 100+ file diffs → summarize top 20 by rank

Refs

Branch refs: .ccs/branches/
File index: .ccs/file-index.md
Project map: .ccs/project-map.md
Task template: .claude/skills/_ccs/templates/task-template.md

10x-Code v2.0.0

diff