Back to skill

Security audit

model-info

Security checks for vulnerabilities and agentic risk

Overview

This skill is a simple model-status helper with disclosed runtime-metadata output, but users should know it uses broad trigger phrases that may activate during ordinary model-related questions.

Install this only if you are comfortable with the agent answering model-status questions using local runtime/session metadata, including provider, API key source/type, endpoint, token usage, and configuration details. Consider invoking it explicitly with "model info" to avoid accidental disclosure during casual conversation.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger phrases are very generic (e.g. "what model", "which AI", "model status") and can easily appear in normal conversation, which may cause the skill to activate unintentionally. While this skill appears informational and not directly privileged, accidental invocation can leak session/runtime metadata the user did not explicitly intend to disclose.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger phrase "what model" is broad enough to match ordinary user conversation rather than an intentional invocation of this skill. In this skill's context, accidental activation is more dangerous because the skill claims it can reveal runtime, provider, session, API key source, and usage details, which could expose sensitive environment metadata unexpectedly.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest lists broad natural-language triggers such as "what model" without clarifying scope, exclusions, or negative examples. Because this is a manifest file, vague invocation phrasing that overlaps with ordinary conversation can make activation conditions unclear.

Content

No source excerpt is available for this finding.