Back to skill

Security audit

Api Tester From Bilibili

Security checks for vulnerabilities and agentic risk

Overview

This API-testing skill is mixed with unrelated content and extra triggers, making it likely to behave unpredictably even though it contains no executable malware.

Review before installing. This skill does not appear to contain malware or high-privilege behavior, but it should be cleaned up to keep only API-testing instructions and narrow triggers; otherwise an agent may load irrelevant guidance for loosely related requests.

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
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Vague Triggers

High
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger phrases are broad enough that normal user requests about APIs, endpoints, or performance could invoke this skill even when the user did not intend to use it. Because the skill content is contaminated with unrelated material, overbroad activation increases the chance of incorrect behavior, context hijacking, or accidental execution of the wrong workflow.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
78% confidence
Finding

The user-facing description, trigger scenarios, and most instructional content are presented in Chinese, while no opt-in or language selection mechanism is documented. This can violate language/locale policy where skills should not impose a specific language without user choice or clear regional justification.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill body contains large blocks of unrelated and contradictory content, including rail transport, hardware teardown, and generic video-learning artifacts that do not align with an API testing skill. In an agent setting, this kind of prompt/data contamination can cause the skill to activate with the wrong context, follow irrelevant instructions, or leak into unintended workflows, reducing predictability and opening a path for prompt injection via merged content.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The 'related help' wording creates an unbounded activation condition that can match loosely associated requests. In a skill already containing unrelated merged content, this ambiguity makes unintended invocation more likely and can route users into irrelevant or unsafe instructions.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This secondary trigger block repeats the same overly broad 'related help' activation pattern, expanding the skill's surface area for accidental selection. Duplicate ambiguous triggers are especially risky here because the document includes multiple embedded sub-skills that do not belong to API testing.

Content

No source excerpt is available for this finding.