Security Audit

avalonia-layout-zafiro

github.com/sickn33/antigravity-awesome-skills

AI SkillCommit e36d6fd3b3f6

CRITICAL

Scanned about 2 months ago

Critical

Immediate action required

High

Priority fixes suggested

Medium

Best practices review

Low

Acknowledged / Tracked

Trust Assessment

avalonia-layout-zafiro received a trust score of 20/100, placing it in the Untrusted category. This skill has significant security findings that require attention before use in production.

SkillShield's automated analysis identified 3 findings: 1 critical, 1 high, 1 medium, and 0 low severity. Key findings include Prompt Injection Attempt, Potential Data Exfiltration via Absolute Path, Excessive Permissions Declared.

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 20, 2026 (commit e36d6fd3). SkillShield performs automated 4-layer security analysis on AI skills and MCP servers.

Layer Breakdown

Manifest Analysis

100%

Static Code Analysis

78%

Dependency Graph

100%

LLM Behavioral Safety

70%

Behavioral Risk Signals

Filesystem Write

3 findings

Dynamic Code

2 findings

Excessive Permissions

2 findings

Security Findings3

Severity	Finding	Layer	Location
CRITICAL	Prompt Injection Attempt The skill attempts to inject instructions into the host LLM by including directives like 'Read ONLY files relevant to the layout challenge!' within the untrusted content. This tries to manipulate the LLM's behavior and constrain its actions, which should be ignored as per security guidelines. Remove all direct instructions or behavioral directives intended for the LLM from the untrusted skill content. Skill descriptions should be purely informative.	LLM	SKILL.md:10
HIGH	Potential Data Exfiltration via Absolute Path The skill provides an absolute path to a local file, `/mnt/fast/Repos/angor/src/Angor/Avalonia/Angor.Avalonia.sln`, and suggests the LLM 'refer' to it. Given the declared 'Read' permission, this could lead the LLM to access and potentially exfiltrate data from a file outside the skill's intended scope, which might contain sensitive information. Avoid providing absolute paths to local files, especially outside the skill's designated working directory. If examples are needed, they should be self-contained within the skill package or refer to publicly accessible, non-sensitive resources. Review and restrict 'Read' permissions if not strictly necessary for the skill's core function.	Static	SKILL.md:28
MEDIUM	Excessive Permissions Declared The skill declares 'Read, Write, Edit, Glob, Grep' permissions. While the skill's content doesn't explicitly exploit 'Write' or 'Edit' permissions, these are powerful capabilities that allow modification of the filesystem. The combination of these broad permissions with the prompt injection and data exfiltration attempts raises the overall risk profile. Adhere to the principle of least privilege. Only declare permissions that are absolutely necessary for the skill to function. If the skill does not need to modify files, 'Write' and 'Edit' permissions should be removed. 'Glob' and 'Grep' should also be reviewed for necessity.	Static	SKILL.md

Scan History

Embed Code

[![SkillShield](https://skillshield.io/api/v1/badge/25d37a64bcf5fc7f.svg)](https://skillshield.io/report/25d37a64bcf5fc7f)