ClawHub
Back to skill

Security audit

Placement Exclusion Manager

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed marketing workflow for building brand-safety placement exclusions, with a minor scope wording conflict around audience exclusions but no evidence of hidden execution or unsafe behavior.

Before installing, confirm that you want this skill to save campaign exclusion summaries into local memory and treat audience exclusions as a dependency from audience-segment-builder. The publisher should align the frontmatter wording with the body so agents do not think this skill creates audience suppression lists.

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
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest/description says the skill produces an 'audience/negative-audience exclusion set', while the body later says audience exclusions are owned by another skill and must only be referenced. This inconsistency can cause an agent to overstep scope, invent or modify suppression lists without the intended source of truth, and generate evidence that misrepresents what was actually reviewed.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The inline instructions conflict on whether the skill 'produces' an audience exclusion set or must not 're-derive' and only reference one from another skill. In an agentic workflow, contradictory delegation boundaries are dangerous because they can lead to inconsistent enforcement, incorrect exclusions, and unreliable A1 audit evidence or handoffs between skills.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.