Your commit history is an unfiltered record of your habits. roast-my-git reads every line of it and tells you the truth.

---

Code reviews catch bad code. Nothing catches bad git habits — until now. roast-my-git walks your full commit history and produces a no-mercy report: when you actually commit (2am?), how conventional your messages are, your longest streak, your biggest blast-radius commit, your branch naming style, and how often you reach for --force.

  roast-my-git
  Analyzing nick@example.com

  ────────────────────────────────────────────────────────────
  Total commits        347
  Commits per week     4.2
  Longest streak       12 days
  Active since         847 days

  Timing
  Most active hour     11pm  (night owl detected)
  Most active day      Wednesday
  Night commits        42  (12%)  — you live dangerously
  Weekend commits      31  (9%)
  Friday PM commits    8   (2%)   — classic

  Message Quality
  Conventional         61%  — room to grow
  Short messages       23   (7%)  — "wip", "fix", "x" count here
  Avg message length   38 chars

  Commit Size
  Average size         84 lines/commit
  Biggest single       1,204 lines  — that's a war crime

  Style
  Branch naming        kebab-case
  Merge commits        18%
  Force pushes         3   — you monster

  ────────────────────────────────────────────────────────────

Install

No npm account needed — runs straight from GitHub:

npx github:NickCirv/roast-my-git

Run from inside any git repository.

Usage

# Analyze your own commits (uses git config user.email)
npx github:NickCirv/roast-my-git

# Analyze a specific author
npx github:NickCirv/roast-my-git --author nick@example.com

# Analyze the whole team (top 10 contributors)
npx github:NickCirv/roast-my-git --team

Flag	Description	Default
--author <email>	Analyze a specific author by email	git config user.email
--team	Analyze all authors (top 10 by commit count)	off

What it analyzes

Category	Metrics
Timing	Hour-by-hour breakdown, most active hour and day, night commits (10pm–4am), morning / afternoon / evening counts
Weekend habits	Weekend commit count and percentage
Danger zones	Friday afternoon commits (2pm–6pm), Monday cleanup commits
Message quality	Conventional commit compliance (feat:, fix:, etc.), average message length, short message count (≤4 chars)
Streaks	Longest consecutive-day commit streak
Commit size	Average lines changed per commit, biggest single-commit blast radius
Style	Branch naming convention (kebab-case / snake_case / camelCase), merge commit ratio, force push count

What it is NOT

• Not a code quality tool. It reads commit metadata and git history — it never touches your source files or diffs for logic.
• Not a productivity score. More commits is not better. The report surfaces patterns; interpretation is yours.
• Not a surveillance tool. It reads your local git log — nothing leaves your machine.

---

_{Node 18+ · MIT · by NickCirv}