Back to skill

Security audit

Deep Loop Thinker

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a local Chinese-language reasoning helper, but it asks agents to record user identity and future needs after every use without clear consent, storage, or deletion limits.

Install only if you are comfortable with a Chinese-language reasoning skill and configure your agent not to store per-user identity, sensitive problem details, or future needs unless the user explicitly consents and you have a defined retention/deletion policy.

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
Findings (1)

other

Note
Location
SKILL.md:52
Finding

Unnecessary Recurring Collection of User Information

Content
View full analysis
Remediation
View remediation
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 (7)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly instructs collecting and recording user feedback along with identity-related details such as whether the user is anonymous or specifically described, but provides no privacy notice, minimization guidance, retention limit, or handling safeguards. In an agent setting, this can normalize unnecessary collection of personal data and lead to unintended storage, disclosure, or reuse of sensitive user information.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill uses plain-language instructions to persistently record user feedback and profile-like details, which creates a real data retention risk even without code. Because agents may operationalize such instructions literally, this can result in unnecessary storage of personal or sensitive information and increase exposure through logs, prompts, or downstream tooling.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The listed triggers such as emotional fluctuation, repeated indecision, or recurring problems are common situations that can arise in many ordinary conversations. The file does not define explicit invocation phrases, scope boundaries, or negative examples, so the activation criteria are too vague for reliable use.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The module description and usage instructions are entirely in Chinese, and the skill presents itself as a general-purpose deep-thinking tool rather than a region-specific or Chinese-only skill. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The runtime interface shown to users uses only Chinese strings such as the title, question prompt, and menu options. This enforces a specific language during operation without offering an alternative or explaining a required locale context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

Natural-language instructions, headings, and examples are entirely in Chinese, which can amount to an implicit language policy for users. The file does not state that the skill is Chinese-only, region-specific, or that users may opt into another language.

Content

No source excerpt is available for this finding.

Ssd 3

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The repeated instruction to record user feedback for iteration reinforces ongoing retention behavior and increases the likelihood that an agent or operator will store user-supplied data by default. While less severe than the detailed collection template, it still promotes persistent data handling without safeguards, consent, or minimization.

Content

No source excerpt is available for this finding.