捷帮文件处理

• Outcome: Review
• Latest audit: Jun 19, 2026 · 10:32 AM UTC
• Version: 1.0.0

Overview

This skill performs the advertised hash and encoding tasks, but it sends user-provided text or file content to an external API using an admin-style environment credential.

Install only if you are comfortable with the content you process being sent to jiebang.site and with the skill using an admin-style API credential. Avoid using it for secrets, tokens, private URLs, proprietary files, or personal data unless the publisher adds clear data-handling disclosure, confirmation before upload, and least-privilege credentials.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (11)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 89% confidence
• Finding: The skill advertises simple text transformations but apparently uses network and environment-variable access without declaring those capabilities. Hidden access to external services and secrets increases the trust boundary and can expose user data or credentials unexpectedly, especially for operations that could be performed locally.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 96% confidence
• Finding: There is a clear mismatch between the declared functionality and the actual behavior: user-provided content is sent to a third-party API and an environment-sourced admin key is used for authentication. This is dangerous because users may provide sensitive text assuming local processing, while the skill silently exfiltrates it to an external service and expands risk through privileged secret handling.

Description-Behavior Mismatch

Medium

• Confidence: 97% confidence
• Finding: The skill claims to perform hash/file-processing style transformations, but the implementation sends the user-provided plaintext to an external service at jiebang.site. This creates a confidentiality and trust-boundary issue because users may reasonably expect local processing for sensitive text, hashes, or file verification inputs, yet their data is transmitted off-platform without disclosure.

Description-Behavior Mismatch

Medium

• Confidence: 96% confidence
• Finding: Base64 encoding/decoding is a trivial local operation, but the code forwards the full user input to a remote API. This unnecessarily exposes potentially sensitive content and misrepresents the privacy properties of the skill, since users would normally expect such processing to occur locally.

Description-Behavior Mismatch

Medium

• Confidence: 96% confidence
• Finding: URL encoding/decoding is also a simple local transformation, yet the implementation sends arbitrary user input to an external service. This creates an avoidable privacy risk and violates user expectations for a low-complexity utility that should not require third-party processing.

Context-Inappropriate Capability

Medium

• Confidence: 92% confidence
• Finding: The skill retrieves an admin-scoped credential and uses it for routine user-driven text transformations against a remote API. Using a high-privilege key for low-risk operations increases blast radius if the environment, logs, downstream service, or code path is compromised, and it suggests over-privileged design.

Intent-Code Divergence

High

• Confidence: 99% confidence
• Finding: The CLI dispatcher passes args.action, which contains the subcommand name ('base64' or 'url'), instead of the requested encode/decode option to the handlers. This breaks intended behavior and can cause incorrect requests to the backend, potentially producing errors or unexpected server-side handling of malformed action values.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill omits a warning that user input is sent to an external API for processing. For a utility handling arbitrary text or file content, this can lead users to unknowingly submit secrets, credentials, personal data, or proprietary material to a third party.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: User-supplied text is transmitted to a remote API for hashing without any user-facing warning, despite hashing often being used on sensitive material such as file contents, tokens, or verification strings. The lack of disclosure prevents informed consent and increases the likelihood of accidental exfiltration of confidential data.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The code sends user content to a remote API for Base64 processing without warning the user. Because Base64 is frequently applied to secrets, binary blobs, or tokens, silent transmission to a third party can expose sensitive data unnecessarily.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The skill silently sends arbitrary user data to a remote service for URL encoding/decoding, even though this operation can be performed locally. Without disclosure, users may unknowingly share confidential URLs, query strings, tokens, or embedded credentials with an external domain.

Static analysis

No suspicious patterns detected.