Security audit

Aight Utils

Security checks for vulnerabilities and agentic risk

Overview

This skill is mostly aligned with an Aight reminders/tasks integration, but it needs review because it can make persistent status changes from broad phrases and may save failed items into local memory files.

Review before installing if you expect the skill to write only to Aight. It should be tightened to require explicit confirmation for ambiguous updates, use the user's actual timezone, and avoid or ask permission before saving reminder/task content into local memory files.

A.I.G

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

SkillSpector

Vulnerability Patterns

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
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The skill advertises broad activation conditions like reminders, deadlines, tasks, or tracking, which overlap with common conversational language and can cause the agent to invoke a state-changing integration too eagerly. In context, this is dangerous because the skill writes into the user's Aight app, so accidental activation can create or modify records without sufficiently clear intent confirmation.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The description explains what the skill does but does not clearly warn that it performs persistent writes and status updates in the user's Aight app. In a write-capable integration, lack of explicit disclosure weakens informed consent and makes accidental data creation or modification more likely, especially when paired with broad trigger rules.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The listed trigger phrases include ambiguous language such as 'Done,' 'Cancel,' and generic task/reminder wording that could appear in ordinary conversation unrelated to Aight operations. Because the skill can create or update persistent items, insufficiently scoped phrases increase the chance of unauthorized or mistaken state changes in the user's task system.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill hard-codes Asia/Shanghai time interpretation for natural-language dates without checking the user's actual locale or asking for confirmation. For a scheduling/reminder skill, this can silently create reminders and deadlines at the wrong times, causing missed deadlines, premature triggers, or erroneous status workflows in a persistent system.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The documented fallback behavior expands the skill from a narrowly scoped Aight integration into writing local logs and memory files, which materially increases its authority and data-handling surface. If triggered by routine failures, the skill could persist sensitive reminder/task content into unintended local files, creating privacy, integrity, and cross-skill data leakage risks beyond the user's expected destination.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The documentation says deadline creation should use trigger as the create type, and the rules later state reminders should use type: "trigger" with scheduledFor. But the deadline example shows two type fields, including type: "deadline", which contradicts the earlier type model and would change the item's semantic type rather than just its category.
Content: No source excerpt is available for this finding.