ClawHub
Back to skill

Security audit

Second Order Effects

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local decision-analysis aid that persistently stores decision notes and preferences, with no code execution or external data transfer found.

Install only if you are comfortable with a local folder of decision analyses and preferences being kept under ~/second-order-effects/. Review or delete memory.md, patterns.md, and decision archives periodically if they contain sensitive, outdated, or unwanted information.

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
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • 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)

Memory Manipulation

High
Category
Memory Poisoning
Content
## Notes

- Save incrementally - don't wait for "complete"
- Decision files are append-only (never delete history)
- Update patterns.md when you notice recurring chains
- Use Decision Index to quickly find past analyses
Confidence
90% confidence
Finding
The instruction that decision files are append-only and history should never be deleted promotes indefinite retention of user decision records, outcomes, and inferred patterns. For a memory-oriented skill, this increases the danger because it can preserve sensitive behavioral profiles and outdated or harmful inferences with no correction, minimization, or deletion path.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The template explicitly instructs creation of a persistent memory file under the user's home directory and encourages ongoing accumulation of preferences, decisions, and patterns without any privacy notice, consent boundary, retention limit, or handling guidance. In a skill that profiles user decision-making habits, this creates a real risk of unnecessary long-term storage of sensitive personal or strategic information.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The setup explicitly instructs the agent to save every learned preference to memory.md immediately, but it does not require clear user notice or consent before persisting data. This creates a privacy and data-governance risk because users may disclose sensitive preferences or personal context during ordinary conversation without realizing it will be stored.

Ssd 3

Medium
Confidence
97% confidence
Finding
The skill encourages conversational harvesting of domains, risk tolerance, stakeholders, and past regrets, then directs that every learned preference be persisted. Even though framed as personalization, this broad retention pattern can capture sensitive personal or strategic information beyond what is necessary, increasing privacy exposure and the blast radius of any later misuse or leakage.

Static analysis

No suspicious patterns detected.