ClawHub
Back to skill

Security audit

Plakar backup

Security checks for vulnerabilities and agentic risk

Overview

This restore workflow is coherent, but it can overwrite live files and is triggered by broad phrases like “undo,” so users should review its scope before installing.

Install only if you want your agent to help with Plakar filesystem recovery. Before any restore, make sure the agent identifies the exact snapshot, path, and destination, and prefer restoring to a separate directory first unless you explicitly want live files overwritten.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill uses broad trigger phrases like 'undo', 'rollback', 'restore', or 'revert', which are common in ordinary conversation and may cause the agent to invoke destructive restore workflows too eagerly. In this context, the skill governs filesystem recovery operations, so accidental activation could steer the agent toward listing snapshots and preparing to overwrite files, increasing the risk of unintended state changes.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The example triggers include ambiguous phrases such as 'undo that' and 'something went wrong, can we go back', which overlap heavily with normal dialogue and could activate a restore skill even when the user means conversational revision rather than filesystem rollback. Because this skill can lead to restoration of live files, ambiguous activation is more dangerous here than in a read-only or informational skill.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The documentation earlier distinguishes between restoring all files to their original paths (restore <snapshotID>) and restoring to a specific directory (restore -to ... <snapshotID>). In the example workflow, step 3 says it is restoring from the chosen snapshot but uses the -to form with <original-path>, which contradicts the documented command semantics and could mislead an agent about what the command actually does.

Content

No source excerpt is available for this finding.