AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified Apache-2.0 Self-run

Conventional Commit Batcher

skill-cnkang-conventional-commit-batcher-conventional-commit-batcher · by cnkang

Auto-split mixed changes into logical commit batches with validated Conventional Commit messages. MUST BE USED for ANY git add, git commit, or git push operation.

No reviews yet
0 installs
11 views
0.0% view→install

Install

$ agentstack add skill-cnkang-conventional-commit-batcher-conventional-commit-batcher

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-cnkang-conventional-commit-batcher-conventional-commit-batcher)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
5mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Conventional Commit Batcher? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Conventional Commit Batcher

Use this skill to turn a messy working tree into clean, reviewable Conventional Commit history.

Commit Interception (MANDATORY)

This skill MUST be activated for ANY commit-related operation, including:

  • user asks to commit, stage, or push changes
  • user asks to "save" or "check in" code
  • any workflow that would result in git add, git commit, or git push

Do NOT run git add or git commit directly without going through this skill's workflow.

Use This Skill When

  • any commit operation is performed through an agent (automatic interception)
  • changes from different intents are mixed in one branch
  • you want a plan-first commit process before opening a PR
  • you need reliable Conventional Commit messages across team and agents

Skip This Skill When

  • the change is tiny and clearly single-intent
  • you only need one quick commit without batching

Note: even when the skill could be skipped, if it is installed, the agent will still run the workflow and safety gates. The result may be a single batch, which is fine.

Default Behavior: Auto-Execute

By default, the skill inspects changes, splits into logical batches, runs safety gates, and commits directly — without waiting for user confirmation.

To see the plan before execution, explicitly ask:

Show me the commit plan first before executing.

High-Success Prompt (Plan-First Mode)

Use this prompt only when you want to review the plan before execution:

Inspect my current git changes and split them into logical Conventional Commit batches.
Output a full Commit Plan first.
Do not run git add or git commit until I confirm the plan.
After confirmation, execute each batch one by one.

Output Contract

In auto-execute mode (default), the agent outputs a brief per-batch summary as each batch is committed.

In plan-first mode (user requested), the agent outputs the full plan before any staging/commit:

Commit Plan
Batch #1: 
Intent: 
Files/Hunks:
-  (...)
Staging commands:
- git add ...
Commit command:
- git commit -m "..."

Required Behavior

  • Always load references/core-rules.md first.
  • Treat references/core-rules.md as the single source of truth.
  • Run python3 scripts/precommit_safety_gate.py before every commit attempt

when Python is available; otherwise run the equivalent manual gate checks from references/core-rules.md.

  • Run the sensitive-data gate before every commit and require explicit user

confirmation if risky files/hunks are detected.

  • When any gate reports risk, include triggered file paths and brief evidence

in user-facing output, plus a concrete "please review" suggestion.

  • Run the .gitignore/local-artifact gate before every commit and require

explicit user confirmation if suspicious files are present.

  • Run branch/conflict/large-file/empty-stage safety checks before every commit.
  • Never skip commit-time checks with --no-verify.
  • If check/hook fails, stop and report concise diagnostics.

Entrypoints

  • Codex skill: SKILL.md
  • Codex repo loader: AGENTS.md
  • Claude: CLAUDE.md, .claude/agents/conventional-commit-batcher.md, .claude/commands/commit-batch.md
  • Kiro: .kiro/agents/conventional-commit-batcher.json, .kiro/steering/commit-batching.md
  • Shared skill (Kimi / Qwen / Gemini): .agents/skills/conventional-commit-batcher/SKILL.md
  • Shared subagent (Qwen / Gemini): .agents/agents/conventional-commit-batcher.md
  • OpenAI: agents/openai.yaml

References

  • Canonical rules: references/core-rules.md
  • Commit plan examples: references/plan-examples.md
  • Commit batching guidance: references/commit-batching-guide.md
  • commit-msg hook example: references/commit-msg-hook-example.md
  • Agent setup docs:
  • references/codex-setup.md
  • references/claude-setup.md
  • references/kiro-setup.md
  • references/kimi-setup.md
  • references/qwen-setup.md
  • references/gemini-setup.md
  • Validator script: scripts/validate_conventional_commit.py
  • Safety gate script: scripts/precommit_safety_gate.py
  • Validator tests: scripts/test_validate_conventional_commit.py
  • Safety gate tests: scripts/test_precommit_safety_gate.py

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.