Security audit

Web Star Studio's Flow PMS

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed FlowDeck project-management API helper, but it can read and change live workspace data using a bearer API key.

Install only if you want an agent to access and modify your FlowDeck workspace. Use a scoped, revocable API key via environment variable when possible, verify the Supabase base URL, and double-check IDs and payloads before delete, archive, broad update, or task-driven implementation actions.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (7)

Description-Behavior Mismatch

Medium

Confidence: 89% confidence
Finding: The script exposes a comments resource even though the stated skill scope is limited to projects, cycles, and tasks. In an agent setting, this scope mismatch increases the chance of unintended data access or mutation through an operation the user and platform may not expect to be available, weakening least-privilege boundaries.

Vague Triggers

Medium

Confidence: 81% confidence
Finding: The trigger text is broad enough to match many generic project-management requests, which may cause the agent to invoke this skill outside clearly intended FlowDeck-specific situations. In security terms, overbroad invocation increases the chance of unnecessary API calls, data exposure, or unintended mutations in a connected workspace.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The file documents permanent delete and archive/unarchive operations without any warning about destructive impact, reversibility, or need for explicit user confirmation. In an agent skill context, omission of such guardrails can lead to accidental or prompt-induced destructive actions against live project data.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: Listing cycle deletion without any cautionary language normalizes a destructive operation and increases the chance an agent or operator will treat it as routine. In project-management tooling, deleting cycles can remove planning history and disrupt reporting, making accidental execution materially harmful.

Missing User Warnings

Low

Confidence: 74% confidence
Finding: The summary states that the API key may be supplied via explicit argument or environment variable but provides no warning about secret handling, logging, or exposure risks. In agent systems, allowing credentials as arguments can increase the chance of accidental disclosure through prompts, traces, or downstream tooling.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: O fluxo instrui a criar uma task via API (`POST /projects/{projectId}/tasks`) com base no pedido do usuário, mas não exige aviso explícito de que haverá uma mutação persistente nem uma confirmação clara antes da escrita. Em um skill que opera sobre um sistema real de gestão, isso pode levar a alterações não intencionais por comandos ambíguos, mal interpretados ou socialmente induzidos.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: A etapa 'Executar o trabalho pedido' liga uma task do Flow à execução prática sem delimitar se isso causará mudanças em arquivos, código ou outros sistemas, nem requer confirmação antes de agir. Esse acoplamento entre leitura de tarefa e ação operacional aumenta o risco de modificações não autorizadas ou excessivas, especialmente se a task contiver instruções amplas, ambíguas ou adversariais.

Static analysis

No suspicious patterns detected.