Back to skill

Security audit

job-tracker

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill benign with medium confidence: scanners found one low-severity inconsistency about a view that may make confirmed changes, while other reported checks found no issue.

Consider whether you are comfortable with the reported inconsistency around confirmed status changes. The scan was partial: although all seven listed components were inspected, a referenced runtime file was missing, so the evidence is incomplete.

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
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (1)

The Ghosted view specifies state changes despite the blanket read-only guarantee.

Low
Category
Excessive Agency
Confidence
98% confidence
Finding

L03 states that all views are read-only and never write the file. However, L12 instructs the Ghosted view to set status_before_ghosted, change status, and append history after user confirmation, explicitly contradicting that guarantee even though the changes are authorized.

Content

No source excerpt is available for this finding.