Back to skill

Security audit

LiblibAI 哩布哩布 libtv 图片生成替代|AI 图片生成与编辑

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly matches its image-generation purpose, but its script trusts service-supplied upload and download details in ways that can affect local files or send selected images to unexpected destinations.

Review this skill before installing in a regular workspace. Use only a dedicated AI Hive API key, avoid sensitive reference images, prefer an isolated Python environment, and keep outputs in a non-sensitive directory. The script should be fixed to validate storage/download domains, reject redirects to unexpected hosts, enforce download size limits, and ensure downloaded files cannot escape the chosen output directory.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/imagegen.py:350
Finding

Unvalidated API Response Enables SSRF, Path Traversal, and Unbounded File Downloads

Content
View full analysis
0: pct = int(downloaded * 100 / total) bar = "█" * (pct // 5) + "░" * (20 - pct // 5) print(f"\r {bar} {pct}% ({downloaded // 1024}KB)", end="", flush=True) ``` ```python task_type = task.get("taskType", "task") for i, item in enumerate(succeeded): result_url = item.get("resultUrl") if not result_url: continue ext = ".mp4" if "video" in task_type.lower() else ".png" filename = f"{task_type}_{task_id}_{i+1}{ext}" out_path = out_dir / filename download_file(result_url, out_path) ``` ### Technical Analysis The application uses two values obtained from the remote task API without validation: 1. `resultUrl` is passed directly to `requests.get()`. No scheme, hostname, port, resolved IP address, or redirect destination is validated. Requests follows redirects by default. A compromised or malicious API response can consequently direct the client to an internal, loopback, link-local, or otherwise unintended HTTP endpoint. 2. `taskType` is included directly in a filesystem path. Path separators, absolute-path syntax, or `..` compon ...[truncated 1924 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/imagegen.py:227
Finding

Reference Images Can Be Uploaded to an Unvalidated Server-Controlled Destination

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Note
Location
SKILL.md:53
Finding

Dependency Installation Instruction Does Not Pin or Verify the Requests Package

Content
View full analysis
Remediation
View remediation
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 (5)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding

The skill advertises executable commands that use shell, file access, environment variables, and network calls, but it does not declare any explicit tool scope or permissions boundaries. This creates a mismatch between the apparent capability surface and the declared contract, increasing the risk of unintended tool invocation, over-privileged execution, or review bypass when the skill is activated.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The description contains broad activation phrases around generic image-generation, editing, migration, and commercial API use, which can cause the skill to trigger for many ordinary user requests beyond the narrow intended context. Over-broad routing is dangerous because it can steer users into a skill with networked/scripted behavior unnecessarily, expanding exposure to external endpoints and local tooling.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
81% confidence
Finding

The skill name and description are written to target Chinese-language search phrases and usage contexts, but there is no statement that users may choose another language or locale. This can create an implicit locale constraint without documented opt-in or justification.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
77% confidence
Finding

The skill's stated purpose is image generation/editing as a LiblibAI alternative, but cmd_init launches a browser to an account page and appends from=cli-skill&skill=... parameters for backend attribution. While credential setup is ancillary, browser-launching with tracking metadata is not an obvious requirement of the image-generation function itself and is not disclosed in the manifest description.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The manifest describes a skill for converting LiblibAI-style workflows into observable visual tasks focused on image generation/editing. In addition to generation, the CLI exposes separate task and upload commands, allowing users to query arbitrary task IDs and upload images independently of a generation request. Those are broader operational capabilities than the user-facing purpose stated in the manifest.

Content

No source excerpt is available for this finding.