Clawhub Gomboc Security Main V0.2.0

• Outcome: Review
• Latest audit: May 29, 2026 · 9:28 PM UTC
• Version: 1.0.0

Overview

This appears to be a real code-remediation skill, but it needs review because it can automate repository changes, commits, and pushes with limited safeguards.

Install only if you are comfortable giving this skill a Gomboc token and allowing automated code-remediation workflows. Start with scan or fix output, use a dedicated least-privilege token, run remediation on a clean branch, review diffs before merge, avoid unattended push workflows, and pin or verify the MCP container before running it.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The README advertises scanning and generating fixes across any codebase but does not clearly warn that remediation actions can modify repository files. In an agent-oriented skill, omission of this side-effect disclosure increases the chance that a user or automated system invokes destructive or unintended code changes without explicit review expectations.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The Quick Start presents a fix command immediately after the scan command without any user-facing warning that it performs code changes. Because this skill is marketed as being for agents and automation, a copied command sequence could lead to unreviewed modifications at scale, making the omission more dangerous than in ordinary documentation.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill explicitly promotes autonomous scanning, remediation, and direct commits/pushes, but does not include a clear warning that these actions can modify repositories and should require explicit user approval or review. In an agent context, this increases the risk of unintended or unsafe code changes being applied automatically, especially because the surrounding language emphasizes trust and autonomy, which may lower operator caution.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The skill instructs users to export a personal access token and later shows a command to display the token, but does not warn that tokens are sensitive secrets that must not be logged, echoed, committed, or exposed in shared shells and CI output. In an agent or automation environment, this omission can lead to credential disclosure through terminal history, screenshots, logs, or unsafe debugging practices.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The workflow example shows an automated remediation path that can generate code changes and create commits/pushes without an explicit warning about the side effects, review expectations, or branch-safety constraints. In CI/CD documentation, this can lead users to enable direct repository modification in trusted workflows without understanding that automation may change code and push updates automatically, increasing the risk of unintended or unsafe changes being merged.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The documentation instructs users to configure a personal access token and invoke scan/fix functionality over an MCP service, but it does not include any guidance on secret handling, transport security, scope minimization, or what code/data may be sent to the service. In an agent integration context, this omission can lead to accidental credential exposure or unintended transmission of sensitive source code to a locally exposed or remotely proxied service.

Missing User Warnings

Medium

• Confidence: 71% confidence
• Finding: The remediate command can trigger remote code changes and optionally commit/push them with minimal user friction and no strong confirmation step. In an agent or automation context, this increases the risk of unintended destructive modifications to local repositories or remote branches if the command is invoked on the wrong path or with unsafe assumptions.

VirusTotal

65/65 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.