ClawHub
Back to skill

Security audit

WorkRally

Security checks across malware telemetry and agentic risk

Overview

The skill appears purpose-aligned for WorkRally, but it needs Review because it grants broad authenticated mutation power with loose in-skill guardrails.

Install only if you intend to let the agent operate your WorkRally account. Review commands before allowing deletes, canvas overwrite, uploads, downloads, or generic `tools call` use, and store the API key only in a protected config directory or environment appropriate for your runtime.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The documented `workrally tools call <tool_name>` interface allows invocation of arbitrary MCP tools beyond the narrowly described media-generation workflows, effectively turning the skill into a generic remote tool dispatcher. In an agent context, this broadens the action surface substantially: if additional MCP tools expose destructive, sensitive, or non-media capabilities, the skill can be steered into unintended operations with only thin documentation-layer guardrails.

Vague Triggers

Medium
Confidence
81% confidence
Finding
The invocation description is extremely broad, covering many common requests such as file upload/download, project management, media generation, and general platform interaction. In an agent-routing setting, overbroad activation criteria can cause the skill to be selected for loosely related prompts, increasing the chance of unnecessary credential use, unintended side effects, or execution of powerful commands without the user clearly intending to invoke this specific external service.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill documents that `auth login` persists the API key to a config file on disk, but it does not prominently warn about local secret storage, file permissions, multi-user environments, or ephemeral/container leakage patterns. In shared agent runtimes, persisted credentials can be exfiltrated by other processes, accidentally baked into images, or retained longer than the user expects.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The guide explicitly states that canvas writes sync in real time to all online users and later documents delete and overwrite operations, but it does not require confirmation or user-warning steps before destructive or shared-state changes. In an AI agent context, this increases the risk of unintended modification or deletion of collaborative content because an agent may execute commands directly from user prompts without surfacing the blast radius.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.