ClawHub
Back to skill

Security audit

Session Guardian

Security checks for vulnerabilities and agentic risk

Overview

This skill advises when to start a fresh session and prepares handoff text, without running code or switching sessions by itself.

Before installing, be comfortable with the agent occasionally suggesting a new session and summarizing useful carry-forward context. Review any handoff before copying it into a new session, especially if the conversation included secrets, private paths, customer names, or sensitive commands; use a per-project install if you do not want global availability.

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
  • 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 (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The default prompt uses broad natural-language concepts like 'continue here' and 'start a new session,' which overlap with ordinary user/agent conversation and can cause unintentional invocation. Because this skill influences session boundaries and context carryover, accidental triggering could prematurely split work, drop relevant context, or inject handoff behavior when the user did not request it.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
81% confidence
Finding

The file instructs the skill to make the boundary recommendation "in the user's language," but then only provides explicit canned text for English and Chinese. This can create a locale-policy issue by implicitly privileging specific languages without stating how other user languages should be handled or offering an opt-in choice.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The markdown instructs the agent to make the boundary note in the user's language, but then only provides canned text for English and Chinese. This can create a locale-policy concern because it implicitly privileges specific languages without explicitly offering equivalent handling or opt-in for other locales.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.