Back to skill

Security audit

GitHub订阅

Security checks for vulnerabilities and agentic risk

Overview

This GitHub Trending skill is mostly documentation-only, but it asks for broad command and file authority that does not fit its narrow public-feed purpose.

Install only if you are comfortable with a skill that declares command execution and file read/write authority for a simple GitHub Trending feed. Prefer a version that limits itself to public GitHub Trending retrieval, removes private-repo and generic automation claims, and documents exactly what commands or files it may use.

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
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (8)

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The activation guidance says to use the skill for broad efficiency, automation, batch processing, and workflow optimization scenarios rather than only for GitHub Trending retrieval. Overbroad trigger conditions can cause an agent router to select this skill in unrelated contexts, increasing the chance that its read/write/exec capabilities are used outside intended scope. The mismatch between narrow name and broad activation language makes this more dangerous, not less.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

Documenting system command execution as a feature of a GitHub Trending feed is an unjustified privilege expansion. A feed retrieval skill should not need arbitrary command execution, and presenting it as normal increases the chance that an agent invokes shell actions under a benign-looking skill label. Given the available exec tool, this creates a meaningful pathway to host-impacting behavior.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The description explicitly says '支持中文交互' as a built-in interaction mode, while the rest of the document does not present language selection or explain a justified region-specific limitation. This can violate language/locale policy when a skill implicitly forces or defaults to a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The usage examples use vague trigger phrasing and generic data/configuration language rather than narrowly describing trending-feed inputs. Ambiguous invocation guidance increases the probability that the skill is activated for unrelated data-processing tasks, which matters because the skill advertises powerful tools beyond its core purpose. This is primarily a scope-control weakness rather than direct code execution logic, but it still elevates misuse risk.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill is presented as a GitHub Trending feed, but the FAQ claims it can process private repositories. That materially overstates scope and may cause an agent or user to assume the skill is authorized to access sensitive private repo data, expanding trust boundaries beyond the declared purpose. In this context, capability inflation is dangerous because the skill already exposes read/exec/write tools, so misleading docs can encourage inappropriate sensitive-data handling.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest describes a narrow trending-feed integration, but later sections expand the scope to generic file processing, API integration, information retrieval, and command execution. This mismatch can cause an agent-selection system or user to invoke a seemingly harmless feed skill that actually encourages broader host interaction, which is a classic trust-boundary violation. Because the declared tools include exec/write, hidden scope expansion is especially risky.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The markdown describes command execution and file writing but does not provide a clear user-facing warning that these actions can affect the local system or workspace. Lack of disclosure undermines informed consent and can cause users or agent frameworks to treat the skill as low-risk when it is not. In a skill nominally focused on public trending data, undisclosed system-impacting behavior is especially problematic.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Advertising broad file read/write behavior is unrelated to a trending-feed skill and normalizes access to local files without a clear functional need. Even if not directly malicious, this can mislead orchestration systems or users into granting filesystem trust to a skill that should only retrieve public remote data. In this context, unnecessary file capability claims increase the attack surface and make misuse more likely.

Content

No source excerpt is available for this finding.