Back to skill

Security audit

using-harness

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small process-guidance skill with broad activation wording, but it does not request sensitive access, run code, persist changes, or move data.

Install this only if you want many project execution, delivery, and verification tasks to be routed through a formal Harness packaging and verifier-evidence process. It is not a data-access or code-execution risk by itself, but it may be too heavy or intrusive for casual tasks.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (2)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill’s trigger conditions are extremely broad and cover generic activities like execution, verification, and delivery, making it likely to activate for many ordinary tasks. That creates a routing/behavior hijack risk where the agent is steered into this skill’s mandatory process even when the user did not request it, reducing user control and potentially interfering with safer or more appropriate workflows.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The description explicitly says to 'Force Harness mode,' which mandates a specific operating mode without user choice. While not directly malicious, this can override normal agent behavior and user preferences, creating a policy-steering issue where task handling is constrained by the skill author rather than the user’s intent.

Content

No source excerpt is available for this finding.