ClawHub
Back to skill

Security audit

Cpa Agents

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed multi-agent orchestration helper with workspace-changing and model fan-out abilities that are bounded by explicit user-control instructions.

Install only the pinned package version with the stated integrity check, and use branch-fix, saga, and fan-out only for tasks where you are comfortable with workspace edits or sending the task text to the named model providers. Do not include secrets or private file contents in fan-out prompts.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
85% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · SKILL.md (reported line 14)May include surrounding context.

md
This skill can change the current workspace and can send task text to more than one model provider.

- `branch-fix` and `saga` may create branches, write files, and request rollback. Run them only after the user explicitly asks for that repository change in this session. Show the planned steps before applying them.
- Stay in the current workspace. Do not read other skill folders, other repositories, or other session transcripts.
- `fan-out` sends the task text to every named model. Do not put secrets, credentials, tokens, private keys, or private file contents in that text.
- Do not call `fan-out` with an empty `models` list. The library would otherwise send the task to `claude-sonnet-4-20250514` and `gpt-4o`. Ask the user which models to use.
- Do not install a floating package range, enable user-global hooks, spawn background agents, or send cross-session messages from this skill.

Static analysis

No suspicious patterns detected.