Security Audit

Sounder25/Google-Antigravity-Skills-Library:18_pre_action_guard

github.com/Sounder25/Google-Antigravity-Skills-Library

AI SkillCommit 09376edc64e9

TRUSTED

Scanned about 1 month ago

Critical

Immediate action required

High

Priority fixes suggested

Medium

Best practices review

Low

Acknowledged / Tracked

Trust Assessment

Sounder25/Google-Antigravity-Skills-Library:18_pre_action_guard received a trust score of 85/100, placing it in the Mostly Trusted category. This skill has passed most security checks with only minor considerations noted.

SkillShield's automated analysis identified 1 finding: 0 critical, 1 high, 0 medium, and 0 low severity. Key findings include Guard allows critical destructive commands despite stated policy.

The analysis covered 4 layers: Manifest Analysis, Static Code Analysis, Dependency Graph, LLM Behavioral Safety. All layers scored 70 or above, reflecting consistent security practices.

Last analyzed on February 28, 2026 (commit 09376edc). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.

Layer Breakdown

Manifest Analysis

100%

Static Code Analysis

100%

Dependency Graph

100%

LLM Behavioral Safety

85%

Behavioral Risk Signals

Filesystem Write

1 finding

Shell Execution

1 finding

Security Findings1

Severity	Finding	Layer	Location
HIGH	Guard allows critical destructive commands despite stated policy The `SKILL.md` documentation explicitly states under 'Stop Conditions' that a 'CRITICAL risk' should result in `allowed: false`. However, the `scripts/guard_check.ps1` implementation sets `$Allowed = $true` for destructive commands (e.g., `rm file.txt`, `del file.txt`) that are not broad root deletions, even though it correctly assigns them a `CRITICAL` risk level. This discrepancy means an agent relying solely on the `allowed` flag might proceed with a critical, destructive action that the skill's documentation implies should be blocked, undermining the core security function of the guard. To align with the stated policy in `SKILL.md`, change `$Allowed = $true` to `$Allowed = $false` within the `else` block for destructive commands. This ensures that any action categorized as `CRITICAL` risk is automatically blocked by setting `allowed` to `false`. If the intent is to allow focused critical actions with a warning, the `SKILL.md` documentation should be updated to reflect this nuance.	LLM	scripts/guard_check.ps1:60

Scan History

Embed Code

[![SkillShield](https://skillshield.io/api/v1/badge/483ce2c33651347d.svg)](https://skillshield.io/report/483ce2c33651347d)