guard-git

A safety and workflow guard for Git, the version-control tool used to track code changes. It installs checks that block risky commands, protect important branches, enforce a commit format, and detect common secrets.

In plain words
What is it for?
Use it before an agent commits, pushes, resets, cleans, deletes branches, or restores files. It supports several coding-agent tools and requires jq when its hook runs.
Why use it?
It helps prevent lost work, accidental changes to shared branches, badly formatted commits, and credentials being checked into the repository.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/danielvm-git/bigpowers/guard-git
Any agent
npx skills add danielvm-git/bigpowers --skill guard-git
Clone the repo
git clone --depth 1 https://github.com/danielvm-git/bigpowers

Made for: Claude Code, Codex.

Per session 100 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,876 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00100 $0.01876
Opus 5 $0.00050 $0.00938
Sonnet 5 $0.00020 $0.00375
Haiku 4.5 $0.00010 $0.00188

Measured 2d ago against content hash 35d2a345eb6d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

guard-git scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

| Claude Code | `PreToolUse` (Bash) | `.claude/settings.json` or `~/.claude/settings.json` |
.cline/skills/guard-git/SKILL.md · 232 lines

How it starts

The opening of the file, as written. The whole thing — 232 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Guard Git

HARD GATEHARD GATE — Before committing, verify: branch is not main/master, author is correct, git user is configured. Bad commits are hard to fix.

Installs a shared hook that blocks destructive git operations and enforces workflow discipline. Requires jq on the agent's PATH when the hook runs.

What gets blocked/enforced

  • Safety: git push --force, git reset --hard, git clean -f, git branch -D, git checkout ., git restore ..
  • Discipline: Blocks direct commits or pushes to protected branches (main, master) unless GIT_BIGPOWERS_LAND=1 (set only by scripts/land-branch.sh).
  • Allows: git push origin <feature-branch> for backup/CI; solo land push to main only inside land-branch.sh.
  • Standardization: Enforces Conventional Commits for all git commit commands.
  • Secrets: Blocks commits containing common secret patterns (sk-, ghp_, AKIA, xoxb-, -----BEGIN private keys) — see REFERENCE.md.

Quick start

  1. Scope: ask project-only vs global (paths differ per product).
  2. Write the hook bundle from REFERENCE.md into the client's hooks directory.
  3. Run chmod +x on pre-tool-use.sh.
  4. Merge the hook snippet from REFERENCE.md into the right settings file — do not wipe unrelated keys.
  5. Verify with the tests in REFERENCE.md.
Client Mechanism Config
Claude Code PreToolUse (Bash) .claude/settings.json or ~/.claude/settings.json
Cursor / Cursor CLI beforeShellExecution .cursor/hooks.json or ~/.cursor/hooks.json
Gemini CLI BeforeTool + run_shell_command .gemini/settings.json or ~/.gemini/settings.json
Google Antigravity Built-in Terminal Deny list Settings UI (no shell hook)

Modes (env on the hook command): GIT_GUARDRAILS_MODE is claude (default) or cursor → stderr + exit 2 on block. Set gemini for Gemini CLI → JSON decision on stdout.

Read the full file on GitHub · 232 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 232 lines · 100 tokens per session scan B 35d2a345eb6d

Subscribe to this mod's changes

guard-git is a skill published in the GitHub repository danielvm-git/bigpowers (156 stars, last pushed 25d ago), licensed MIT. It adds 100 tokens to every session and 1,876 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.