Security Audit

Jamkris/everything-gemini-code:skills/eval-harness

github.com/Jamkris/everything-gemini-code

AI SkillCommit 6c6f43aaa3d6

CRITICAL

Scanned 8 days ago

Critical

Immediate action required

High

Priority fixes suggested

Medium

Best practices review

Low

Acknowledged / Tracked

Trust Assessment

Jamkris/everything-gemini-code:skills/eval-harness received a trust score of 43/100, placing it in the Untrusted category. This skill has significant security findings that require attention before use in production.

SkillShield's automated analysis identified 2 findings: 0 critical, 2 high, 0 medium, and 0 low severity. Key findings include Potential Command Injection via Bash tool, Excessive Permissions: Broad tool access (Bash, Read, Write, Edit).

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 March 30, 2026 (commit 6c6f43aa). 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

70%

Behavioral Risk Signals

Network Access

2 findings

Filesystem Write

1 finding

Shell Execution

2 findings

Dynamic Code

2 findings

Excessive Permissions

1 finding

Security Findings2

Severity	Finding	Layer	Location
HIGH	Potential Command Injection via Bash tool The skill's manifest explicitly requests the `Bash` tool, and the `SKILL.md` provides examples of `bash` commands (`grep`, `npm test`, `npm run build`) that are likely to be executed by the agent. If user-controlled input (e.g., `feature-name` in `/eval define feature-name` or parts of the 'Code-Based Grader' commands) is directly interpolated into these `bash` commands without proper sanitization, it could lead to arbitrary command execution on the host system. Implement strict input validation and sanitization for any user-provided strings that are used in `bash` commands. Prefer using safer, parameterized execution methods or specific tool functions over direct shell interpolation where possible. Ensure that any arguments passed to `npm` commands are also properly validated.	LLM	SKILL.md:60
HIGH	Excessive Permissions: Broad tool access (Bash, Read, Write, Edit) The skill's manifest requests `Bash`, `Read`, `Write`, and `Edit` tools. While the skill's purpose (evaluating code, running tests) may require some filesystem interaction, the combination of arbitrary command execution (`Bash`) and broad filesystem access (`Read`, `Write`, `Edit`) grants the agent extensive control over the host environment. This significantly increases the attack surface for data exfiltration, command injection, and unauthorized system modifications if the agent is compromised or misused. Review and minimize the requested tools to the absolute minimum necessary for the skill's core functionality. Consider using more granular or sandboxed tools instead of broad `Bash` access. Implement strict access controls and monitoring for actions performed by the agent using these powerful tools.	LLM	Manifest

Scan History

Embed Code

[![SkillShield](https://skillshield.io/api/v1/badge/1d69ad949f5ccdb2.svg)](https://skillshield.io/report/1d69ad949f5ccdb2)