git-version-control
PassAudited by VirusTotal on May 11, 2026.
Overview
Type: OpenClaw Skill Name: git-version-control Version: 1.0.0 thought The user wants me to analyze an OpenClaw AgentSkills skill bundle (ZIP content provided as text). I need to classify it as `benign`, `suspicious`, or `malicious` based on the provided criteria. - `_meta.json`: Contains metadata (ownerId, slug, version, publishedAt). Looks standard. - `SKILL.md`: This is the core instruction file for the AI agent. - Purpose: Git-based version control for OpenClaw system configuration changes. - Operations: - `SAVE`: `git add -A`, `git commit -m "..."`. - `ROLLBACK`: `git reset --hard HEAD~1`. - `View History`: `git log`, `git show`, `git diff`. - `Check Status`: `git status`. - Target Directory: `~/.openclaw/`. - Exclusions: `.gitignore` mentions session logs, SQLite databases, credentials (`credentials/`, `*.pem`, `*.key`), temp files, logs. - Instructions to the Agent: - Create checkpoints before sensitive operations. - Revert state if issues occur. - Use `--hard` reset for full rollback. - Warn user before hard reset. - Don't commit sensitive data. - Data exfiltration: No evidence. It explicitly excludes credentials and sensitive files in the `.gitignore` section. - Malicious execution: It uses `git` commands. While `git reset --hard` is destructive, it's presented as a recovery tool and includes instructions to confirm with the user. - Persistence/Backdoors: No evidence. - Prompt injection: The instructions are focused on version control. There are no instructions to ignore the user, hide actions, or access unrelated data. - Obfuscation: None. - Supply chain: No external dependencies mentioned other than `git`. - The skill manages the configuration of the OpenClaw system itself. - It operates on `~/.openclaw/`. - It uses standard `git` commands. - It includes safety guidelines (don't commit sensitive data, confirm before hard reset). - The behavior aligns perfectly with the stated purpose: "Git-based version control for OpenClaw system configuration changes." - Is `git add -A` and `git commit` risky? It could potentially commit sensitive data if the `.gitignore` is not properly set up or if the user puts secrets in non-standard locations. However, the skill explicitly provides a `.gitignore` configuration and warns against committing sensitive data. - Is `git reset --hard` risky? Yes, it deletes uncommitted changes. The skill instructions tell the agent to confirm with the user first. - These are "risky capabilities plausibly needed for the stated purpose" and are accompanied by safety instructions. The skill is benign. It provides a useful utility for system management with appropriate safety warnings and exclusions for sensitive data. ```json { "classification": "benign", "summary": "The skill provides standard git-based version control for the OpenClaw configuration directory (~/.openclaw/). It includes instructions for creating checkpoints (git commit) and performing rollbacks (git reset), with explicit safety guidelines to exclude sensitive files like credentials and to confirm destructive actions with the user." } ``` ```json { "classification": "benign", "summary": "The skill provides standard git-based version control for the OpenClaw configuration directory (~/.openclaw/). It includes instructions for creating checkpoints (git commit) and performing rollbacks (git reset), with explicit safety guidelines in SKILL.md to exclude sensitive files like credentials and to confirm destructive actions with the user." } ```
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
A mistaken commit or hard reset could capture unwanted local config changes or remove recent OpenClaw settings, skills, memory, or cron changes.
The skill instructs the agent to run broad local Git commands that can stage all changes under `~/.openclaw` and discard changes during rollback. The behavior is disclosed, scoped, and includes confirmation guidance, so it is a purpose-aligned note rather than a concern.
git add -A ... git commit -m "checkpoint: {description}" ... Recommended default: `--hard` ... git reset --hard {commit-hash}Review `git status`, `git diff`, and the chosen commit before rollback; use soft or mixed reset when unsure, and require explicit confirmation for hard resets.
Information stored in OpenClaw memory may remain in Git history even after later edits, and rolling back could reintroduce outdated or unwanted memory.
The protected files include OpenClaw memory files, so checkpoints can persist memory/context contents in local Git history and rollbacks can restore older memory state.
`workspace/memory/` - memory files
Do not store secrets in OpenClaw memory, verify tracked files with `git ls-files`, and clean Git history manually if sensitive data was committed.
The skill may fail or behave inconsistently on systems where Git is not installed or where `~/.openclaw` is not a Git repository.
The skill is documented as instruction-only and declares no binary requirements, while its visible workflow depends on Git commands. This is an under-declared dependency, not evidence of hidden installation or code.
Required binaries (all must exist): none ... No install spec — this is an instruction-only skill.
Declare Git as a required binary and document the expected repository setup before use.