ClawHub
Back to skill

Security audit

Perfect Handoff

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent project handoff helper with a local read-only checker and explicit limits.

Install this only if you want an agent to review project documentation and write or update a compact handoff file. It can read relevant project files and edit documentation when you authorize closeout, so invoke it at a project root you trust and review the produced handoff before relying on it in a fresh session.

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 (1)

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The declared purpose says the skill should produce a project handoff and ready-to-paste prompt for a fresh session. The supplied code does not generate any handoff content. Instead, it reads a specified handoff file and project root, parses markdown headings, enforces required sections, checks size limits and placeholders, verifies prompt/path inclusion, and validates local link existence. This is a materially different primary purpose: validation of an already-written handoff, not reconciliation or authoring. No suspicious extra capability beyond local filesystem reads is present, but the core behavior does not match the declared description.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.