Security audit

Hita-Mind & Knowledge

Security checks for vulnerabilities and agentic risk

Overview

The skill is a local memory and knowledge tool, but it encourages automatic long-term storage of user preferences, contacts, projects, and daily notes without clear consent or retention controls.

Install only if you are comfortable with an agent keeping local long-term memory. Avoid storing secrets, credentials, regulated personal data, customer data, or confidential project details unless you have reviewed where the JSON files are stored and how to inspect or delete entries.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (10)

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The README explicitly describes automatically saving user preferences and task corrections into persistent memory, but provides no notice, consent model, retention policy, or guidance on handling sensitive data. In a memory-oriented agent skill, this creates a real privacy and data-governance risk because users may unknowingly disclose personal preferences, work habits, or corrective feedback that gets stored long-term in local JSON files.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The README lists categories such as projects, daily logs, and contacts, and states storage is via JSON files, but does not warn that potentially sensitive operational or personal information may be retained on disk. This is dangerous because users and operators may deploy the skill without realizing it persists contextual task data and contact information, increasing the chance of unintended disclosure, over-collection, or insecure local storage.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger list includes very broad phrases such as “知识库” and memory-related prompts that can appear in normal conversation, which may cause the skill to activate unintentionally. In a memory/knowledge management skill, unintended activation is risky because it can surface stored memories or knowledge entries, or cause the agent to store conversational content without the user clearly intending to invoke the skill.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The README explicitly describes automatically saving user preferences and corrections into long-term memory, but provides no notice, consent model, retention policy, or privacy safeguards. In an AI memory skill, this can lead to silent persistence of sensitive personal data, instructions, or behavioral profiling information that the user may not expect to be stored long term.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill explicitly states it stores long-term memory including preferences, decisions, contacts, feedback, projects, and daily conversation records, but it provides no privacy notice, consent flow, retention policy, or guidance on handling sensitive data. In an agent-memory context, users may unknowingly cause personal or confidential information to be persisted, increasing the risk of privacy violations, over-collection, and unintended reuse in later interactions.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The documentation says `memory-store.json` is automatically created on first run, but it does not clearly warn users that invoking the skill writes persistent data to local disk. Silent persistence is risky because conversation content and preference data may be stored without the operator realizing it, which can expose sensitive information through local file access, backups, syncing tools, or later agent retrieval.

Missing User Warnings

Medium

Confidence: 86% confidence
Finding: The module automatically creates and persists a JSON memory store on disk without any notice, consent flow, retention policy, or visibility into what is stored. In a memory/knowledge-management skill, this increases privacy risk because users may provide sensitive preferences, contacts, project notes, or daily logs that are silently retained across sessions.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The delete command permanently removes stored entries immediately after matching the ID, with no confirmation, soft-delete, undo, or backup. Because this skill is specifically designed for long-term memory accumulation, accidental or induced deletion can cause irreversible loss of important agent state and user knowledge.

Ssd 3

Medium

Confidence: 90% confidence
Finding: The README instructs the agent to automatically retain user-provided preferences in persistent memory. In the context of an agent memory system, this is especially risky because preferences can include personal, behavioral, or sensitive workflow information, and the automatic nature removes an explicit consent checkpoint before long-term storage.

Ssd 3

Medium

Confidence: 88% confidence
Finding: The README directs the agent to keep daily and project memory for user task handovers and context, which can capture substantial operational details, internal project information, or sensitive customer data. Because this skill is specifically designed for long-term memory accumulation, the context makes the persistence risk more serious than a transient note-taking feature.

Static analysis

No suspicious patterns detected.