Back to skill

Security audit

hermes-model-router

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent model-routing helper that classifies tasks and stores local provider configuration, with no evidence of hidden execution, exfiltration, or privilege escalation.

Installers should be comfortable with a Chinese-language workflow and should review provider settings before using cloud routing, especially for confidential prompts. The local config file may be created or overwritten at ~/.model-router/config.json.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (9)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill explicitly routes complex tasks to third-party cloud providers but does not warn users that their prompts, potentially including sensitive data, may leave the local environment. This creates a real privacy and data-governance risk because users may assume a 'local-first' router keeps all task content on-device unless clearly told otherwise.

Content

No source excerpt is available for this finding.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
80% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · install.sh (reported line 51)May include surrounding context.

sh
# Create config directory
mkdir -p "$HOME/.model-router"
chmod 700 "$HOME/.model-router"

echo ""
echo -e "${GREEN}✅${NC} Skill installed to: $TARGET_DIR"

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This code's natural-language interface and descriptions are entirely in Chinese, including the module description and all user-facing CLI output, with no indication that other languages are supported. The policy specifically calls for flagging language or locale constraints when the skill forces a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The script's natural-language title, prompts, and guidance are all in Chinese, which imposes a specific language on users. There is no indication that the user can select another language or that the skill is intentionally restricted to a Chinese-speaking context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The natural-language instructions and usage examples are presented in Chinese, which can amount to a language/locale constraint if the skill is expected for general use. The file does not indicate that Chinese is optional, nor does it justify the locale restriction as region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The package description is written entirely in Chinese, and the file provides no indication that language is selectable or that the skill is intentionally limited to Chinese-speaking users. This may violate language/locale policy because it imposes a specific language without documented opt-in or justification.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
87% confidence
Finding

This manifest description says the skill routes 'simple tasks' and 'complex tasks' but does not define what kinds of requests should invoke the skill or where its scope begins and ends. In a manifest file, this kind of broad capability phrasing can create ambiguous activation conditions and unintended invocation overlap with many ordinary AI tasks.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill description under the manifest metadata is entirely in Chinese, with no alternative language text or note that the skill is region- or locale-specific. That creates a language constraint not clearly offered as a user choice.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

This code persists configuration data to ~/.model-router/config.json, which is a file write affecting user data on disk. Although it prints a success message after saving, there is no disclosure before the write explaining that the wizard will create or overwrite a local config file.

Content

No source excerpt is available for this finding.