ClawHub
Back to skill

Security audit

TokenKiller

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed token-saving workflow guide with some broad activation and formatting defaults, but no evidence of hidden code, data access, persistence, or unsafe behavior.

Install this if you want agents to be more concise and conservative with context. Be aware that it may activate during ordinary debugging, large-repo, cost, or multi-step requests; explicitly ask for full logs, full files, or more detail when completeness matters more than token savings.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Vague Triggers

Medium
Confidence
80% confidence
Finding
This markdown file states that the skill includes 'Trigger words for auto-enable' but does not identify the trigger phrases, constraints, or exclusion conditions in this file. That lack of specificity can indicate ambiguous activation behavior and makes it hard to assess whether common speech could unintentionally invoke the skill.

Vague Triggers

Medium
Confidence
96% confidence
Finding
The auto-enable trigger list is broad enough to activate on common user phrases such as cost, long logs, large repos, or multi-step work, which can silently alter agent behavior without clear user consent. In this skill, that means compression and gating rules may be applied in many ordinary tasks, potentially suppressing useful context or causing important evidence to be omitted during debugging, review, or security-sensitive workflows.

Natural-Language Policy Violations

Low
Confidence
85% confidence
Finding
The default output template mandates fixed English section headers like 'Conclusion', 'Evidence', 'Changes/Actions', and 'Next Step' with no opt-in or locale guidance. This can conflict with language/locale policies when users operate in another language and the skill does not state that the template should follow the user's language.

Static analysis

No suspicious patterns detected.