ワンクリックで
repo-guard
Automatically detect and prevent dangerous operations, secret exposure, and main branch modifications.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Automatically detect and prevent dangerous operations, secret exposure, and main branch modifications.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | repo-guard |
| description | Automatically detect and prevent dangerous operations, secret exposure, and main branch modifications. |
| allowed-tools | ["Read","Grep","Glob","Bash(git:*)","Bash(rg:*)","Bash(ls:*)"] |
| context | inline |
Proactively protect the repository from common mistakes and security issues.
This skill activates automatically when:
Detect and block commits containing:
Patterns to watch:
- .env files
- *_KEY, *_SECRET, *_TOKEN patterns
- BEGIN PRIVATE KEY
- password=, passwd=, pwd=
- Bearer tokens
- AWS credentials
Action: Block and suggest using environment variables or secret managers.
Block or require confirmation for:
rm -rf /
rm -rf *
rm -rf .
mkfs.*
dd if=
> /dev/sd*
chmod -R 777
Action: Block immediately with warning, suggest safer alternatives.
Prevent direct modifications to:
Action: Suggest creating feature branch first.
Warn about:
Action: Warn and suggest .gitignore or Git LFS.
**Repo Guard Alert**
**Issue Detected**: [Type of issue]
**Location**: [File or command]
**Severity**: [Critical/Warning/Info]
**Details**:
[Explanation of the issue]
**Recommended Action**:
[What to do instead]
**To proceed anyway** (not recommended):
[Explicit override instructions]
No output - allow operation to proceed silently.
Automatically invoked before:
Referenced in rules for file modification warnings.
Included in system prompt for command validation.