Skeall Skill Builder

• Outcome: Review
• Latest audit: May 28, 2026, 1:50 AM
• Version:
  1.0.0

Overview

Skeall is a coherent skill-building and auditing helper, but its healthcheck mode can inspect local agent configs, environment variables, session logs, and URLs with loose boundaries.

Install only if you want an agent to audit and edit skill files. Use create/scan/improve on explicit paths, review diffs before accepting changes, and avoid running healthcheck-all unless you are comfortable with local agent configs, session logs, environment-variable names, and discovered URLs being inspected in the active session.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The trigger phrase 'any SKILL.md question' is overly broad and can cause the skill to activate for many unrelated requests involving SKILL.md files. Over-broad activation increases the chance that the agent applies this skill in the wrong context, which can override more appropriate instructions and expand the skill's influence unexpectedly.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: Generic triggers like 'build a skill', 'create skill', and 'improve skill' are broad enough to match many benign or unrelated user requests. This can lead to unintended invocation, prompt-routing confusion, and increased attack surface because the skill may become active in sessions where it was not specifically desired.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The template explicitly tells authors to include generic trigger phrases like "any {trigger-1}, {trigger-2}, or {trigger-3} question," which encourages overly broad routing criteria. In an agent skill system, this can cause unintended invocation on loosely related prompts, increasing the chance the wrong skill influences behavior or overrides more appropriate instructions.

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The checklist enforces the presence of trigger phrases but does not require them to be precise or bounded, so a skill can pass review while still being matched too broadly. In this skill-building context, that omission is significant because it systematizes unsafe routing patterns across newly created skills rather than being a one-off wording issue.

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The example natural-language trigger, "help me with X," is overly broad and resembles ordinary user language rather than a narrowly scoped invocation. In a skill system that routes based on trigger phrases, this can cause accidental activation on unrelated prompts, leading to prompt hijacking of normal conversations, incorrect tool routing, or unintended disclosure of the skill’s behavior in contexts where it should remain inactive.

VirusTotal

64/64 vendors flagged this skill as clean.