Back to skill

Security audit

repo-orientation

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because it reports instructions to change file permissions during a workflow described as read-only, though the available evidence does not show malicious intent.

Consider the suspicious verdict before installing. The reported instruction could change executable permissions and affect how files are treated; Skillspector recommends removing automatic permission changes and reporting mode differences instead. Its analysis was partial because a local path-like reference in SKILL.md could not be resolved, so the review 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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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
Findings (2)

The skill instructs the agent to change file permissions despite explicitly promising read-only inspection.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The instruction to run chmod +x before inspecting differences contradicts the skill's explicit read-only boundary and can cause unauthorized changes to executable permissions. Although presented as correcting Windows synchronization artifacts, it assumes permission differences are accidental and may obscure legitimate security-relevant changes or make untrusted files executable; it does not itself execute those files.

Content

No source excerpt is available for this finding.

The read-only inspection workflow instructs the agent to change executable permissions without warning or user approval.

Medium
Category
Output Handling
Confidence
99% confidence
Finding

This is the same underlying issue as SDI-4: a supposedly read-only orientation workflow directs the agent to modify permissions without obtaining approval. The surrounding prohibitions on writes reduce the apparent intended scope, but the conflicting concrete command creates a credible path to unauthorized changes and misleading reports that omit the original permission differences.

Content

No source excerpt is available for this finding.