Effective Git

• Outcome: Review
• Latest audit: May 29, 2026 · 9:27 PM UTC
• Version: 1.0.0

Overview

This is a real Git helper, but some quick commands are described as low-risk while they can change branches, stash state, commits, and local repository contents.

Review the specific gq command before using it. Treat quick commands as potentially mutating, inspect status and diffs before quick commits or amends, avoid running the diff shortcut in repositories with untracked secrets, and confirm pushes, rebases, resets, and conflict resolutions explicitly.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (11)

Intent-Code Divergence

Medium

• Confidence: 96% confidence
• Finding: The README labels quick operations as 'read-only' or low-risk, but the documented `gq` commands include branch switching, branch creation/deletion, stash save/pop, and committing, all of which modify repository state. This mismatch can cause users or higher-level agents to authorize commands under a false safety assumption, increasing the chance of unintended changes or destructive actions.

Intent-Code Divergence

High

• Confidence: 94% confidence
• Finding: The conflict-resolution section first states the assistant must never automatically choose one side, then later presents 'Keep Ours' and 'Keep Theirs' as resolution strategies. This contradictory guidance can justify silent loss of code during merge or rebase conflicts, especially if an agent interprets the later section as permission to auto-resolve without user review.

Intent-Code Divergence

Medium

• Confidence: 93% confidence
• Finding: The document states that quick commands are 'read-only or low-risk operations,' but the reference includes clearly state-changing actions such as branch switching, branch creation/deletion, stashing, stash pop/apply, fetch, and commit/amend. This mismatch can cause users or downstream agents to trust the shortcut path as safer than it is, increasing the chance of unintended repository changes without appropriate confirmation or review.

Intent-Code Divergence

Medium

• Confidence: 90% confidence
• Finding: The help text for the diff command says it shows unstaged changes, but the implementation also prints full contents of untracked files. That mismatch can cause users to expose secrets or sensitive new files they did not expect to be displayed, paged, or sent to another terminal process.

Vague Triggers

Medium

• Confidence: 80% confidence
• Finding: The natural-language activation examples such as 'help me commit code' and similar workflow phrases are broad enough that normal conversation could trigger repository-modifying behavior. In an assistant skill that can commit, rebase, merge, and push, ambiguous activation increases the risk of accidental execution of sensitive Git actions.

Vague Triggers

Medium

• Confidence: 80% confidence
• Finding: The manifest description is broad enough to trigger on many generic Git-related requests, which can cause the skill to activate in contexts where the user only wanted advice rather than repository-affecting actions. In a tool-capable Git assistant, over-broad activation increases the chance of unintended command execution and makes other unsafe behaviors more dangerous.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: Natural-language triggers such as general phrases about committing code or resolving conflicts overlap with ordinary conversation and can activate the skill without sufficiently explicit user intent. Because this skill can lead to repository-modifying operations, ambiguous activation creates a meaningful risk of unintended Git actions.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The document instructs use of `git reset --hard B-before-rebase`, which is a destructive command, but the nearby text does not explicitly warn that `--hard` discards uncommitted working tree and index changes. In a Git-assistant skill, users may follow commands mechanically during a stressful conflict/rebase situation, increasing the chance of accidental local data loss.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The quick-command reference documents destructive or mutating operations such as branch deletion and stash pop without prominent warnings about side effects, reversibility, or prerequisites. In an agent skill context, concise shortcut docs can encourage execution with reduced scrutiny, making accidental data loss or workflow disruption more likely.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The script executes programs taken directly from environment variables, including an unquoted terminal command string and a pager pipeline. If a user runs this script in a hostile or modified environment, attacker-controlled values can trigger arbitrary command execution under the user's account.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The quick commit command performs `git add -A` before committing, which stages all tracked, deleted, and untracked changes. In a fast-path git helper, that can easily commit secrets, temporary files, unrelated work, or generated artifacts that the user did not intend to include.

VirusTotal

64/64 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.