Security audit

Linear流程CLI(专业版)

Security checks across malware telemetry and agentic risk

Overview

This skill is mostly coherent for Linear automation, but it gives an agent broad ability to change project data and integrations with limited guardrails.

Review this before installing if your Linear workspace contains sensitive or business-critical data. Use suggest-only or preview-required mode by default, restrict the Linear token to the minimum needed permissions, verify Slack/Ticket content before sending it to Linear, and only enable autonomous mode, Git hooks, webhook changes, or bulk deletes in a controlled workspace with trusted URLs and rollback procedures.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (5)

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The skill repeatedly describes itself as an automation hub that can 'take over' Linear workflows, but it does not define clear activation boundaries, approval gates, or negative examples of when not to run. In an agent runtime, broad capability framing increases the chance of unintended invocation and unsafe write actions against project-management data.

Vague Triggers

High

Confidence: 94% confidence
Finding: The applicable-keywords list includes very broad generic terms such as automation, workflow, efficiency, issue, and agent runtime. In systems that route or auto-select skills by keyword matching, this can cause accidental selection of a powerful exec/write-enabled skill for loosely related prompts, increasing the chance of unintended command execution or external API writes.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The Slack/Ticket integration promotes parsing message content and automatically creating or triaging issues, but it does not clearly warn that Slack or ticket text may contain sensitive business, personal, or security information that will be transmitted into Linear. In an agent context, this can normalize silent cross-system data movement and create privacy, compliance, and over-sharing risks.

Missing User Warnings

Medium

Confidence: 89% confidence
Finding: The webhook section shows creation, update, and deletion of outbound integrations without warning that webhooks can exfiltrate issue data to third-party URLs and that deletion is destructive. This is especially risky in an agent-operated environment where an LLM could modify notification infrastructure or send data to attacker-controlled endpoints.

Missing User Warnings

Low

Confidence: 80% confidence
Finding: The Git/JJ automation section encourages automatic issue-state changes based on commit and merge events, but it omits a warning that repository activity can mutate external project state without a separate approval step. This can lead to unintended status changes, especially from malformed commit messages, shared branches, or automated merges.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.