ClawHub
Back to skill

Security audit

What Did You Do

Security checks for vulnerabilities and agentic risk

Overview

This skill transparently summarizes recent agent actions for the current user and does not install code, persist data, contact external services, or hide extra behavior.

Install this if you want an easy way to ask the agent what it recently did. Be aware that replies may include filenames, command names, and workflow details from the current session, although the skill instructs the agent to redact secrets and avoid file contents.

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
95% confidence
Finding

The skill is configured to activate on broad natural-language phrases such as 'asks what the agent did', 'wants to review recent actions', or 'asks for a log'. Because the skill reveals a reconstructed history of tool usage, broad triggering increases the chance it runs in contexts where the user did not explicitly request operational telemetry, potentially exposing sensitive filenames, commands, or workflow details from the current session.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The prompt "What have you been doing?" is a broad, natural-language trigger that overlaps with ordinary conversation. In a skill that exposes recent agent actions, such an ambiguous activation phrase can cause unintended invocation and disclosure of session activity when the user did not explicitly mean to call the skill.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The prompt "Can you show me what actions you took?" is also an ambiguous conversational phrase that may appear in normal dialogue, increasing the chance that the skill activates unintentionally. Because this skill reconstructs tool calls, targets, and decisions, accidental triggering can leak operational details, filenames, or other sensitive context from the current session.

Content

No source excerpt is available for this finding.