Security audit

ai-kujiale-design-free

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed Kujiale interior-design workflow, but users should handle its local access token carefully and keep command use narrowly scoped.

Install only if you intend to use Kujiale design generation. Keep the access token out of version control and logs, restrict the config file's permissions, rotate the token if exposed, and do not let the agent run unrelated shell commands or process unrelated local files under this skill.

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 (5)

Context-Inappropriate Capability

Medium

Confidence: 83% confidence
Finding: The skill documentation advertises broad file-processing capability unrelated to the narrowly scoped interior-design workflow. In an agent environment with `read`/`write` tools, unnecessarily broad capability claims can normalize or justify access to arbitrary files, increasing the risk of data exposure or unintended file manipulation beyond least-privilege expectations.

Context-Inappropriate Capability

Medium

Confidence: 87% confidence
Finding: The documentation explicitly promotes generic command-execution capability even though the skill's business purpose is a constrained design workflow. In an agent that has `exec`, this can encourage execution of commands outside the necessary API wrapper scripts, expanding the attack surface for command misuse, environment inspection, or unsafe shell interactions.

Context-Inappropriate Capability

Medium

Confidence: 82% confidence
Finding: This repeated feature section again frames the skill as having general file-processing capability unrelated to its stated function. Repetition reinforces an overly broad operational scope, which is dangerous in agentic contexts because it can legitimize unnecessary file access and weaken reviewer or runtime assumptions about bounded behavior.

Vague Triggers

Medium

Confidence: 79% confidence
Finding: The trigger condition activates when a user merely mentions interior design or renovation, which is overly broad for an agent skill that can read, write, and execute commands. In context, broad auto-triggering increases the chance of the skill engaging during casual conversation and reaching credential-loading or command-execution paths without sufficiently explicit user intent.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill tells users to store a live access token in a local JSON file but does not provide strong warnings about secrets at rest, file permissions, rotation, or safer alternatives. In this agent context, local secret files are particularly sensitive because the same skill also has `read`, `write`, and `exec` capabilities, raising the risk of credential disclosure or accidental inclusion in logs or version control.

Static analysis

No suspicious patterns detected.