ClawHub
Back to skill

Security audit

Collaboration Helper

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed local task tracker that stores shared task notes in a JSON file and does not show hidden or destructive behavior.

Install only if you are comfortable with a shared local task file persisting owners, priorities, and notes across runs. Do not put secrets, credentials, private customer data, or sensitive internal notes into task entries, and use any alternate data path only for task files you intend this helper to read or update.

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
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Context Leakage

High
Category
Data Exfiltration
Content
`scripts/collaboration_helper.py` is a lightweight JSON-backed tracker for community action items:

- `list` shows every open or in-progress task with owner, priority, and creation timestamp.
- `add` creates a new entry using `--owner`, `--priority`, and optional `--note` fields to capture context.
- `complete` marks a task as finished and records who closed it.

The data lives in `data/tasks.json`, so the collaboration state survives across skill runs and reboots.
Confidence
78% confidence
Finding
The skill is designed to persist collaboration notes, owners, and handoff details across runs in a JSON file, which can expose sensitive operational context to later users or agents. In this context, 'capture context' is not dangerous by itself, but the persistent shared task store and support for pointing at another workspace make unintended disclosure of internal notes, priorities, or coordination state plausible.

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill advertises shell execution plus file read/write behavior but does not declare any explicit tool scope or permissions in the manifest. That creates an authorization gap: an agent or reviewer cannot tell from the skill metadata what filesystem or command execution boundaries are intended, increasing the chance of over-privileged execution and unsafe use of the optional `--workspace` path.

subprocess module call

Medium
Category
Dangerous Code Execution
Content
def run_cli(data_path: Path, args):
    return subprocess.run(
        [sys.executable, str(SCRIPT_PATH), "--data", str(data_path)] + args,
        capture_output=True,
        text=True,
Confidence
70% confidence
Finding
subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Static analysis

No suspicious patterns detected.