Security audit

reimagine-it

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed design-transformation assistant that reads project context and writes redesigned artifacts, with no hidden execution, exfiltration, or deceptive behavior found.

Install only if you want an agent to inspect project files and create or modify design artifacts. For existing app or document targets, review the destination path and diff before accepting in-place edits, and use plan-only or companion-output workflows when you do not want source files changed.

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (10)

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The skill advertises very broad natural-language triggers such as 'reimagine it', 'reinvent this', and 'redesign this page', which are plausible in ordinary conversation and can cause accidental invocation. Unintended activation is risky because this skill is empowered to inspect workspace files, infer context, and generate or modify artifacts without an explicit high-friction confirmation step.

Missing User Warnings

Medium

Confidence: 87% confidence
Finding: The guidance explicitly allows writing output in place when an existing landing page is present, but it does not require confirmation, backup creation, or any warning before overwriting user files. In an agent skill, this creates a real risk of destructive or unintended modification of project artifacts, especially if the agent applies the instruction automatically from a vague redesign request.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The routing logic explicitly authorizes the skill to choose actions from broad contextual cues like vague user intent, repository state, or inferred needs. In an agent setting, this can cause overreach: the skill may perform materially different file edits or create new artifacts without a clear, narrowly scoped user instruction, increasing the chance of unintended modification, confusion, or misuse of host capabilities.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The activation rule is broad enough that the skill may load for common terms like 'document', 'docx', or 'md' rather than an explicit user request for this specific redesign behavior. In an agent setting, overly permissive triggering can cause unintended execution of file-reading and transformation workflows, increasing the chance of surprise actions or misuse in contexts where the user did not mean to invoke the skill.

Missing User Warnings

Low

Confidence: 84% confidence
Finding: The skill instructs the agent to write output files into the workspace without any explicit user-facing warning or confirmation step. While the path appears scoped to a reimagined subdirectory, silent file creation can still surprise users, overwrite expected artifacts if naming collides, or normalize agent-side modification without consent.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: This section explicitly instructs writing generated PDFs under the workspace without requiring user confirmation or a prior warning. In agentic environments, silent file creation can lead to unexpected repository changes, overwriting of existing artifacts, or creation of files that get committed or consumed by later automated steps.

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The skill explicitly allows `/reimagine-it simulation <any words the user typed>`, which broadens activation and passes arbitrary trailing text into the skill's 'open brief' behavior. In an agent setting, this can cause unintended invocation on ambiguous requests and let untrusted residual prompt text steer generation in ways the router or user did not clearly intend.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger guidance says to load this form when the user 'forces' common terms like 'slides', 'pptx', or 'deck'. These are broad everyday words that can appear in many unrelated requests, making unintended activation likely and causing the agent to apply this skill in contexts the user did not explicitly request.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The activation guidance explicitly loads this universal form when the user forces `universal` or when the router does not recognize a file, making it a broad fallback for arbitrary inputs. In an agent skill context, that increases the chance the skill will operate on unexpected or unsupported file types and then proceed into file inspection, toolchain selection, and output generation paths that may be unsafe or error-prone for untrusted content.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The activation rule is intentionally expansive, triggering not only on explicit form-router choices but also broadly whenever the context merely 'is a page' such as an existing index.html, docs page, landing page, or dashboard. In an agent setting, this can cause the skill to load in situations the user did not clearly request, increasing the chance that strong opinionated transformation instructions override task scope and lead to unintended file modifications or unsafe autonomous behavior.

Static analysis

No suspicious patterns detected.