Back to skill

Security audit

automation-repair-triage

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious with medium confidence because two scanners flag instructions to retry rejected operations through another interface, which could bypass approval controls.

Review the interface-switching instructions before installing. In particular, check that the skill stops on permission, policy, or approval denials instead of retrying through another path. The scan is partial because referenced memory files were missing, so the evidence about persistence 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
Findings (1)

T01 · Skill Instruction Hijacking

Error
Location
.
Finding

T01: Skill Instruction Hijacking

Content
View full analysis

Impact

The agent is instructed to attempt operations through an alternate interface after access is denied. This can undermine tool-specific approval controls where another execution interface remains available; successful circumvention depends on the host's enforcement.

Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Instructs the agent to route around rejected operations by changing interfaces.

High
Category
Prompt Injection
Confidence
97% confidence
Finding

The skill explicitly directs the agent to change interfaces after an operation is rejected, including switching between tools and CLI commands, without distinguishing authorization denials from ordinary technical failures. This can turn a security rejection into a cue to bypass the enforcing interface and perform the same operation through a less restricted path. Although the examples concern mostly read-only diagnostics, which limits immediate harm, the unrestricted fallback rule could undermine approval controls or expose restricted logs and system information.

Content

No source excerpt is available for this finding.