Security audit

Deep Reading

Security checks for vulnerabilities and agentic risk

Overview

This is a Chinese deep-reading note-taking skill whose local file writes, optional quality checker, memory integration, and cron template are disclosed and fit the stated reading workflow.

Install this if you want a Chinese close-reading workflow that creates persistent reading notes. Review the reading/ and memory file behavior before use, and enable the cron template or companion memory skills only for texts and notes you are comfortable storing and updating automatically.

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

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (10)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 98% confidence
Finding: The declared purpose is a reading assistant, but the skill also introduces local file inspection and quality-check workflow behavior that is not clearly reflected in the top-level description. Description-behavior mismatch is dangerous because users may consent to a harmless-seeming reading tool without realizing it can read files and drive validation workflows, which weakens informed consent and increases misuse risk.
Content: No source excerpt is available for this finding.

Undeclared Tool Scope

Medium

Category: MCP Least Privilege
Confidence: 87% confidence
Finding: The skill instructs the agent to read local files such as reading/plan.md and book content, but it does not declare any explicit tool scope or permissions boundary. That creates a least-privilege and transparency problem: users and hosting platforms cannot clearly see what file access the skill expects, increasing the chance of unintended local file reads or overbroad runtime access.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The trigger list includes broad everyday phrases like '看书' and '读书笔记', which can cause accidental activation in normal conversation. In a skill that may read or create files, unintended activation increases the chance of surprising file operations, misrouted user requests, or autonomous workflow initiation without clear intent.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The skill description and all operational instructions are written as if the assistant should operate in Chinese, but the document does not offer a language choice or state that the locale restriction is intentional and limited to a Chinese-only use case. This can violate language/locale policy because it effectively imposes a specific language without user opt-in.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The file structure and workflow describe automatic generation and updating of local files, but the skill does not prominently warn users that local content will be created or modified. Silent or poorly disclosed writes are dangerous because they can overwrite user work, create unwanted artifacts, or store sensitive material without informed consent.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 80% confidence
Finding: The skill is positioned as a reading assistant, yet it is also designed to feed long-term memory synchronization and co-occurrence systems. Even if intended as a productivity feature, this broadens the data flow and retention surface beyond what a user may expect from simple reading notes, increasing the risk of unintended persistence or propagation of sensitive text excerpts.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The cron workflow expands the skill from chapter note generation into broader memory-file updates and insight-pipeline participation. Scheduled autonomous writes are more sensitive than interactive note generation because they can modify user data repeatedly without fresh review, creating integrity and privacy risks if the plan file or prompts are wrong or manipulated.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The cron instructions encourage scheduled autonomous execution that reads plans, writes notes, updates progress, and appends memory entries, but they do not clearly warn that these file modifications will occur unattended. Unattended recurring writes materially increase the risk of data corruption, privacy exposure, or propagation of bad outputs from a stale or manipulated plan.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The file is entirely written as a Chinese-only workflow and prescribes fixed user-facing messages in Chinese, such as the required notification and reply phrases. This creates a language/locale constraint without documenting user choice or opt-in, which matches the policy-violation category for forced language behavior.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: This Python file contains natural-language instructions and output text exclusively in Chinese, including the module docstring and CLI usage guidance. Under the policy rule for language/locale, forcing a specific language without user opt-in or documented justification is a violation.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.