Back to skill

Security audit

todo-manager

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because it reports an undeclared health check and automatic saving of TODO findings, though the scan was only partial.

Consider postponing installation until the health check is explained and saving requires a clear choice. The scan was partial: only 2 of 4 components were fully scanned, and some behavior and referenced paths could not be assessed.

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
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (2)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The TODO scan and summary align with the description. However, importing quick-healthcheck and calling healthcheck.quick() introduces an undeclared diagnostic capability that is not visibly needed to scan comments. The supplied chunk does not reveal the health check's internals, so no specific resource access or side effects can be established.

Content

No source excerpt is available for this finding.

Task-file synchronization is automatic rather than optional when TODOs are found.

Low
Category
Excessive Agency
Confidence
92% confidence
Finding

The CLI calls saveTodos whenever matches exist, despite persistence being described as optional and the documented CLI offering no explicit save choice. This creates an unexpected filesystem side effect and can persist source comments and context that the user intended only to inspect; the unseen saveTodos implementation determines overwrite behavior and exact storage risks.

Content

No source excerpt is available for this finding.