Clear Mind

• Outcome: Review
• Latest audit: Sep 12, 2026 · 1:13 AM UTC
• Version: 0.1.0

Overview

The skill is a coherent memory-cleanup tool, but it can rewrite and delete persistent memory containing sensitive details without sufficiently explicit backup and approval controls.

Review this skill carefully before installing. It does not show exfiltration, remote code, or deception, but users should only run it when they intentionally want long-term memory reorganized, and should require a verified backup plus explicit approval before any migration or deletion of portfolio, user directive, project, or incident records.

A.I.G

Vulnerability Patterns

• Insecure Skill Coding Practices: Finds exploitable flaws such as hardcoded secrets or command injection
• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally

Findings (1)

T09 · Insecure Skill Coding Practices

Warning

• Location: SKILL.md:93
• Finding: Sensitive Information Migrated to Predictable Plaintext Files Without Mandatory Pre-Migration Backup## Vulnerability Details **File Location**: `SKILL.md`, lines 93-115; related deletion and backup instructions at lines 50-53 and 175-179 **Vulnerability Type**: Plaintext sensitive-data storage and unsafe destructive migration **Risk Level**: Medium ### Vulnerable Code ```markdown Ensure `/memory/facts/` exists with these files: ``` memory/facts/ ├── critical-events.md # Important incidents and lessons ├── projects.md # Active project statuses ├── technical-setup.md # DEPRECATED — technical configs now stay in MEMORY.md ├── portfolio.md # Stocks, investments, strategies └── user-directives.md # User preferences and instructions ``` **Note**: `technical-setup.md` is kept for backward compatibility. New technical content goes to MEMORY.md. ### Phase 3: Migrate Content Move factual details from MEMORY.md to appropriate facts/ files: | Source Section | Target File | Migration Rule | |---------------------|----------------------|-----------------------------------------------------------------------------------------------| | Critical Events | critical-events.md | Move full details, keep index link | | Projects | projects.md | Move all project details | | Technical Setup | technical-setup.md | **KEEP in MEMORY.md** — agent capabilities stay; only move niche/deprecated configs if requested | | Portfolio/Trading | portfolio.md | Move holdings, strategies | | User Directives | user-directives.md | Move preferences, quotes | ``` Related destructive behavior: ```markdown ### DELET ...[truncated 3923 chars]
• Remediation: ## Remediation Suggestions 1. Require explicit, informed approval immediately before migrating each sensitive category, especially financial records, personal directives, and historical incidents. 2. Create a versioned backup of `MEMORY.md` and the complete `facts/` directory before every write, move, overwrite, or deletion operation. 3. Validate the backup by checking readability, file counts, and cryptographic hashes before proceeding. 4. Use an atomic migration transaction: write to a private temporary directory, validate content completeness, atomically rename files into place, and roll back automatically on any failure. 5. Prohibit deletion during automatic maintenance. Present proposed deletions to the user and require explicit confirmation. 6. Apply least-privilege file permissions to `MEMORY.md`, `facts/`, temporary files, and backups. Avoid shared or globally readable directories. 7. Detect and redact secrets such as credentials, tokens, private keys, account identifiers, and authentication data rather than moving them into additional plaintext files. 8. Minimize duplicated sensitive information and define retention periods for migrated records and backups. 9. Perform post-migration integrity checks to verify that all retained content is represented exactly once and that core rules, capabilities, and user directives remain intact. 10. Record an auditable migration manifest containing source section, destination, timestamp, checksum, and user authorization without logging the sensitive content itself.

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The activation criteria include vague triggers like 'Memory cleanup is requested' and 'Periodic maintenance,' which can cause the skill to run without a clearly bounded need or scope. In a memory-management skill that rewrites and migrates user data, ambiguous invocation conditions increase the chance of unnecessary modification, over-collection, or accidental movement/deletion of important information.

Whitespace Padding

Medium

• Category: Prompt Injection

Content

Move factual details from MEMORY.md to appropriate facts/ files:

| Source Section      | Target File          | Migration Rule                                                                                |
|---------------------|----------------------|-----------------------------------------------------------------------------------------------|
| Critical Events     | critical-events.md   | Move full details, keep index link                                                            |
| Projects            | projects.md          | Move all project details                                                                      |

• Confidence: 70% confidence
• Finding: Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Static analysis

No suspicious patterns detected.