ClawHub
Back to skill

Security audit

Knowledge Graph - Transitive Closure Generator

Security checks for vulnerabilities and agentic risk

Overview

The skill is a local graph-analysis utility with no obvious data-theft or persistence behavior, but it materially overstates implemented cycle-detection and algorithm support.

Review before installing if you need reliable dependency or hierarchy validation. The skill appears locally contained, but do not rely on its advertised cycle detection, Floyd-Warshall, incremental, memoization, or materialization behavior without fixing or independently validating the implementation.

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 (4)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The skill advertises cycle detection, multiple selectable algorithms, and materialization behavior that the finding says are not actually implemented. In a reasoning or dependency-analysis context, this can mislead downstream agents or users into trusting incomplete graph analysis, causing missed cycles, incorrect reachability assumptions, and unsafe decisions based on false completeness. The skill context increases risk because transitive closure is often used for dependency, hierarchy, and policy reasoning where silent inaccuracies can propagate broadly.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The enum advertises FLOYD_WARSHALL and INCREMENTAL algorithms, but the dispatcher silently falls back to BFS for unsupported options. This creates deceptive behavior where callers may believe they selected a specific algorithm with expected correctness or performance properties, leading to integrity and availability issues under large or specialized workloads.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The class/config documentation implies cycle detection support, but the code path does not perform that security-relevant validation. In this skill context—dependency analysis, hierarchies, and reachability—misleading documentation increases the chance that other components rely on nonexistent safeguards and process hostile or malformed graph input unsafely.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The configuration and skill description claim cycle detection support, but the implementation never checks for cycles and always leaves has_cycle as False. In dependency analysis or hierarchy validation workflows, this can cause consumers to trust incorrect acyclic assumptions, potentially bypassing validation logic or producing unsafe downstream decisions.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.