Commit Message Generator

Critical Rules

Never ask for confirmation

When this skill is invoked, the user wants to commit now. Do NOT ask "shall I commit?" or "ready to proceed?" -- analyze, generate the message, and commit immediately.

Never bypass pre-commit hooks

• •FORBIDDEN: --no-verify, -n, noqa, skip-test, skip-ruff
• •CORRECT: git commit -S -m "message" (hooks run normally)
• •If hooks fail, fix the issue instead of bypassing

Never add CoAuthor

No Co-authored-by lines for any AI agent.

Always prevent pager blocking

Prefix every git command with GIT_PAGER=cat:

GIT_PAGER=cat git diff --cached --name-status
GIT_PAGER=cat git diff --cached
GIT_PAGER=cat git log --oneline -5

Workflow

Step 1: Check staged files

GIT_PAGER=cat git diff --cached --name-status
GIT_PAGER=cat git diff --cached --stat

If nothing is staged, stop immediately:

Nothing is staged for commit.

Stage changes first:
  git add <file>     # specific files
  git add -A         # all changes
  git add -p         # interactive staging

Step 2: Analyze changes

Read the actual diff:

GIT_PAGER=cat git diff --cached

For each staged file:

1. •Categorize: new feature, bug fix, refactor, test, docs, config, dependency
2. •Identify patterns: related changes across files, domain/component being modified, breaking changes

Step 3: Generate commit message

Format:

<type>(<scope>): <short summary>

<detailed description>

Changes:
- <bullet point 1>
- <bullet point 2>

Impact:
- <impact description>

Files Modified:
- <file 1>
- <file 2>

Types: feat, fix, refactor, test, docs, style, chore, perf

Scope: domain/component name (e.g., inference, allocation, etl). Use * for cross-cutting changes.

Rules:

• •Summary line under 72 characters
• •Imperative mood ("add feature" not "added feature")
• •Be specific about what changed, not just where
• •Highlight breaking changes prominently

Step 4: Commit immediately (no confirmation)

Do NOT ask the user for confirmation. Execute the commit directly.

Only signed commits:

git commit -S -m "<message>"

Display success confirmation with the commit hash.

Example Output

feat(allocation): add HRP risk-parity allocation endpoint

Implemented hierarchical risk parity allocation using covariance matrix
input, producing portfolio weights that respect risk-parity constraints.

Changes:
- Created HRP core function with quasi-diagonal clustering
- Added POST /allocation/hrp endpoint with weight response schema
- Added integration tests calling the endpoint with sample covariance data

Impact:
- Enables HRP as baseline allocator for comparison against PPO/SAC

Files Modified:
brain_api/
├── core/hrp.py (new)
├── routes/allocation.py (new)
└── tests/test_allocation_hrp.py (new)

Statistics: 3 files changed, 280 insertions