ClawHub
Back to skill

Security audit

founder-trajectory-matrix

Security checks for vulnerabilities and agentic risk

Overview

The coaching workflow is coherent, but it directs the agent to retain real-session observations in the reusable skill without requiring consent or removal of private details.

Before installing, remove or revise the instruction to save observations after every session. Require explicit permission to retain examples, remove identifying and confidential details, and review any changes before sharing the skill. The artifacts do not establish malicious intent or automatic disclosure.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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)

The skill instructs agents to retain real coaching failure patterns without a privacy warning.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The instruction to add entries after each real use and 'paste the actual failure pattern' encourages persisting information from potentially confidential founder-coaching conversations without requiring consent, anonymization, or data minimization. Because the destination is a reusable skill associated with an open-source repository, identifiable founder behavior, team conflicts, or business details could reach subsequent users or public contributions. The file does not itself perform a write or upload, so exposure depends on an agent following this instruction and subsequently storing or sharing the entries.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.