Security audit

agent-error-logger-new

Security checks across malware telemetry and agentic risk

Overview

This is mostly a local error-log skill, but it includes GitHub publishing material that can publicly push files and unsafe token guidance, so users should review it before installing.

Install only if you want durable local error logs in workspace/memory. Do not store secrets or sensitive customer/business data in the logs. Do not run create-repo.sh unless you intentionally want to create a public GitHub repository and push the current directory. Avoid the PAT-in-URL example; use GitHub CLI, SSH, or a credential manager instead.

SkillSpector

By NVIDIA

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 Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection

Findings (7)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 96% confidence
Finding: The skill clearly describes persistent storage under workspace/memory and provides commands that write logs and reports, but it declares no permissions. Undeclared file read/write capability weakens the trust boundary for the agent and can lead to unauthorized access or modification of local data when the skill is invoked.

Description-Behavior Mismatch

Medium

Confidence: 88% confidence
Finding: The script creates and publishes a GitHub repository even though the skill is described as an error-logging tool, which introduces an unrelated external side effect. This can cause unintended code publication and repository creation from the user's authenticated GitHub account, expanding risk beyond the skill's stated purpose.

Context-Inappropriate Capability

Medium

Confidence: 90% confidence
Finding: The file uses authenticated GitHub CLI operations to manage the user's GitHub account and repositories, which is not justified by the core error-logging functionality. In skill ecosystems, unrelated account-management capabilities are dangerous because they can perform impactful actions under the user's identity if invoked automatically or without scrutiny.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The document instructs users to embed a GitHub Personal Access Token directly in a `git push` URL. This is dangerous because credentials placed in command lines and repository URLs can be exposed through shell history, process listings, terminal logs, screenshots, or accidental commits to git config/remotes. In the context of a skill-distribution workflow, this increases the chance that non-expert users will leak a reusable GitHub credential.

Vague Triggers

Medium

Confidence: 87% confidence
Finding: The trigger phrases include very generic terms like '记录错误', '查询错误', '错误分析', and '避免踩坑', which can match routine conversation. Because the skill performs persistent logging and retrieval, broad auto-activation increases the chance of capturing or acting on sensitive conversation content without clear user intent.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The command creates a public repository and pushes the current local source with --push, but the script does not present a clear explicit warning or confirmation about immediate public publication. This can leak source code, embedded secrets, internal files, or work-in-progress content from the local directory to GitHub under the user's account.

Ssd 3

Medium

Confidence: 94% confidence
Finding: The skill explicitly offers to extract error details from conversations and store them in persistent markdown logs. Error reports often contain credentials, file paths, tokens, internal URLs, personal data, or business-sensitive context, so automatic extraction into plain-language logs creates a real data retention and disclosure risk.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.