Back to skill

Security audit

ADHD Assistant

Security checks for vulnerabilities and agentic risk

Overview

The skill is purpose-aligned, but it asks to persist sensitive ADHD and treatment information without clear consent or deletion controls.

Review this before installing if you do not want ADHD status, treatment context, routines, triggers, or productivity patterns stored across sessions. Use it only with clear memory settings, and avoid saving medical or mental-health details unless you have explicit control over storage and deletion.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:129
Finding

Persistent Storage of Sensitive Health Information Without Adequate Privacy Controls

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 129-143; related storage instructions at lines 367-371
Vulnerability Type: Sensitive data persistence without explicit consent, minimization, retention, or deletion controls
Risk Level: Medium

Vulnerable Skill Instructions

markdown
Over time, remember these preferences (via SkillBoss API Hub memory):

**Schedule & Energy:**
- Peak focus hours (morning person vs. night owl)
- Typical energy patterns throughout the day
- Best times for deep work vs. shallow tasks

**Task Management:**
- Preferred number of daily priorities (1-3 recommended)
- Task/note storage location (files, apps, directories)
- Preferred reminder frequency and channels

**ADHD Profile:**
- Diagnosed or suspected ADHD
- Current treatments (medication, therapy) - for context only

The related memory integration instructions state:

markdown
### Memory Usage:
- Store user preferences and patterns
- Remember what strategies have worked
- Track routine adherence over time
- Note energy patterns and triggers

Technical Analysis

The Skill explicitly directs the Agent to retain whether a user has diagnosed or suspected ADHD and details about current medication or therapy in the SkillBoss API Hub memory system. These fields constitute sensitive health information.

The instructions do not require informed opt-in consent before persistence, identify a retention period, provide a deletion mechanism, limit which integrations may retrieve the information, or require data minimization. Treatment information is also not necessary for most of the Skill's stated planning and productivity functions. Although the Skill says the information is for context only, that qualification does not establish a technical access or retention control.

This is classified as T09: Insecure Skill Coding Practices because the Skill configuration promotes unsafe handling of ...[truncated 1984 chars]

Remediation
View remediation

Remediation Suggestions

  1. Remove diagnosis status, medication, and therapy details from the default list of information to remember.
  2. Keep health disclosures session-local unless persistent storage is strictly necessary for a user-requested feature.
  3. Require explicit, informed, field-specific opt-in before storing any health information. The prompt should identify what will be stored, why it is needed, where it will be stored, and how long it will remain.
  4. Apply data minimization. Prefer non-medical operational preferences such as desired reminder frequency or preferred focus-block duration.
  5. Define a short retention period and automatically expire sensitive records.
  6. Provide commands or documented procedures that allow users to inspect, correct, and delete all stored information.
  7. Restrict sensitive records from being retrieved by unrelated skills, channels, or integrations.
  8. Avoid placing health information in plaintext logs, task files, reminder messages, or broadly accessible memory namespaces.
  9. Add an instruction requiring the Agent to ask before recalling sensitive information in a new conversation or communication channel.
  10. Document the external memory provider's access controls, encryption expectations, retention behavior, and deletion guarantees.
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 (6)

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
80% confidence
Finding

Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Content

Scanner excerpt · SKILL.md (reported line 50)May include surrounding context.

md
- Provides virtual body doubling sessions
- Creates structured co-working check-ins
- Sets up accountability partnerships
- Offers presence-based support without judgment

### 6. Dopamine Regulation
- Helps build personalized "dopamine menus"

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The top-level description activates on very broad, common productivity and overwhelm cues, so the skill could be invoked for many ordinary planning conversations unrelated to ADHD. In a mental-health-adjacent skill, accidental activation increases the chance of collecting sensitive context or framing normal difficulties as ADHD-related when the user did not explicitly want that mode of support.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The listed activation conditions are ambiguous and overlap heavily with generic productivity, stress, and organization requests. Because the skill is tailored to ADHD and emotional regulation, this ambiguity can cause overreach into sensitive mental-health framing and memory use outside clearly intended contexts.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger phrases are ordinary statements like 'I need help planning' and 'I feel overwhelmed,' which are too generic to safely distinguish this specialized skill from normal assistant behavior. This raises the risk of accidental activation, inappropriate personalization, and unnecessary collection of sensitive behavioral or mental-health-adjacent data.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill proposes remembering diagnosed or suspected ADHD status, treatments, pitfalls, sensitivities, and patterns, which are sensitive mental-health and health-related data. Storing that information without a clear user-facing notice, purpose limitation, and explicit consent creates privacy and data-handling risk, especially because the data could persist across sessions and influence future behavior.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The file-management section authorizes reading and writing task lists and related documents but does not instruct the agent to obtain clear confirmation before accessing or modifying user files. That can lead to unintended data exposure, overwrites, or silent persistence of sensitive productivity and mental-health-adjacent notes.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.