Security audit

Folder UI Visualizer - show the folder through HTML (via Telegram)

Security checks for vulnerabilities and agentic risk

Overview

This skill can list local folder metadata, upload the generated report to Telegram, and delete the local copy while its privacy note incorrectly says no third-party upload occurs.

Review before installing. Use it only if you are comfortable sending folder names, paths, file sizes, and modification dates to Telegram, and avoid running it on sensitive directories. The skill should be revised to make Telegram export opt-in, correct the privacy statement, skip or constrain symlinks, and add traversal limits.

A.I.G

Vulnerability Patterns

Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
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

Findings (3)

T05 · Unauthorized Access and Privilege Escalation

Error

Location: file_lister.js:35
Finding: Symbolic-Link Traversal Escapes the Requested Directory Boundary
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: file_lister.js:26
Finding: Unbounded Recursive Directory Traversal Can Exhaust Process Resources
Content: View full analysis
Remediation: View remediation
MAX_DEPTH) { return []; } } ``` 2. Enforce a global maximum number of visited entries and stop processing when the limit is reached. 3. Set a maximum report size and avoid constructing an unlimited HTML string in memory. 4. Use `fs.lstatSync()` and reject symbolic links by default. 5. Maintain a `Set` of canonical directory paths and skip directories that have already been visited. 6. Consider an iterative traversal algorithm to avoid JavaScript call-stack exhaustion. 7. Apply process-level execution-time, memory, and output-disk quotas. 8. Generate output in a controlled temporary directory and use guaranteed cleanup logic, such as `try`/`finally`, rather than relying exclusively on a later Agent instruction. 9. Return a clear error when a safety limit is reached instead of silently producing an incomplete or unbounded report. ]]>

other

Note

Location: SKILL.md:87
Finding: Privacy Documentation Incorrectly Claims That No Data Is Uploaded to Third Parties
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (11)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 97% confidence
Finding: The skill’s declared purpose and security claims do not match its documented behavior. It presents itself as a local visualization utility with strong security properties, but the instructions include creating a file, transmitting it externally via Telegram, and deleting local evidence afterward; this mismatch can mislead users and agents into authorizing unexpected data handling.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

High

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The skill states that no data is uploaded to third-party servers, yet it explicitly instructs the agent to POST the generated report to the Telegram API. This false privacy claim can cause users or downstream systems to trust the skill under incorrect assumptions, enabling unintended disclosure of local directory contents.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: Requiring a Telegram bot token gives the skill external transmission capability that is unnecessary for a basic folder visualizer. In context, this increases the risk that sensitive filesystem metadata will be sent off-host without the user understanding that the skill is acting as a networked export tool.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The documented workflow goes beyond local visualization by uploading the generated folder report to Telegram and then deleting it. That creates a stealthy exfiltration-and-cleanup pattern, which is risky because directory listings often contain sensitive filenames, project structure, and secrets-related paths.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The user-facing description does not clearly warn that folder contents will be transmitted to Telegram and that the generated file will be deleted afterward. This omission undermines informed consent and increases the chance of sensitive project information being shared unexpectedly.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The trigger phrases are broad enough to match ordinary requests about viewing folder contents, while the skill performs additional sensitive actions like external upload and deletion. This makes accidental invocation more likely in benign contexts where the user expected only a local listing.
Content: No source excerpt is available for this finding.

External Transmission

Medium

Category: Data Exfiltration
Confidence: 93% confidence
Finding: The skill contains an explicit external network transmission path to the Telegram API. In this context, the issue is not merely that networking exists, but that a local directory report containing potentially sensitive filenames is uploaded to a third-party service as part of the default workflow.
Content: Scanner excerpt · SKILL.md (reported line 72)May include surrounding context.

BOT_TOKEN (string): The Telegram bot token. Read this from the $TELEGRAM_BOT_TOKEN environment variable. Example:

bash
curl -s -X POST "https://api.telegram.org/bot<BOT_TOKEN>/sendDocument" -F "chat_id=12345678" -F "document=@C:\Users\Username\Desktop\file.zip"

Step 3 (Cleanup): REGARDLESS of success, immediately delete the file.

Windows: del "<path>"

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: Using path.resolve() only converts the supplied path to an absolute path; it does not constrain access to a trusted root. If this skill is exposed through an agent or automation context, an attacker can supply arbitrary absolute or relative paths and cause the tool to enumerate sensitive local directories and metadata outside the intended scope.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 87% confidence
Finding: The manifest emphasizes creating a collapsible HTML visualization of local folders, which suggests generating a representation of directory contents. In implementation, the skill persists that visualization as a timestamped HTML file on the local filesystem, which is an additional write operation not stated in the description.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 77% confidence
Finding: The file contains user-relevant natural language in Chinese comments and English runtime/output text, but does not provide any language or locale selection or document that it is intended for a specific audience. Under the stated policy, forcing a language/locale without opt-in can be a violation.
Content: No source excerpt is available for this finding.

Missing User Warnings

Low

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The script writes an HTML file to the current working directory and only emits the resulting path afterward. While the operation is visible in code, there is no prior confirmation prompt or user-facing warning in comments/output near execution to disclose that a new file will be created.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.