Back to skill

Security audit

Bilibili video

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious with medium confidence, mainly because it found a concern about whether the reported capabilities fit the skill’s purpose, and parts of the artifact were only partially inspected.

The scan recommends against installing, and the artifact was only 45% fully inspected. Consider waiting for a more complete inspection and for the reported publishing success check to be addressed; the verdict is suspicious, not a claim that the author acted maliciously.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 49)May include surrounding context.

md
`ego-browser nodejs` 不继承 shell 环境变量,也不转发 argv,所以由 `publish.mjs` 把 JSON 写到 `/tmp` 再交给 ego。**参数全部走 flag**(key=value 也支持):

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 140)May include surrounding context.

md
`post-video.mjs` 报错后,JSON 里的 `spaceId` 可以回到原 space 排查,不要另开 space:

Success detection promises a URL-change check plus matching text, but accepts matching text alone (L1164, L1181).

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The success loop accepts any page text matching PUBLISH_OK_RE without enforcing the documented URL-change condition or requiring a submission-specific receipt. Matching stale or attacker-influenced text could therefore produce a false success report, misleading downstream automation; the regex definition is not provided, so the exact spoofing inputs cannot be confirmed.

Content

No source excerpt is available for this finding.

The documented plan-only mode claims to check login, but exits without invoking the browser workflow.

Low
Category
Excessive Agency
Confidence
95% confidence
Finding

The comment at L003 promises "login-check + plan only" without --yes. The plan branch only prints the plan and exits with code 3, before the browser workflow at L095, so it does not perform the promised login check and can give callers a misleading impression that authentication was validated.

Content

No source excerpt is available for this finding.