Back to skill

Security audit

commit

Security checks for vulnerabilities and agentic risk

Overview

This skill clearly discloses automatic commits and pushes, but its push commands can publish more branches than promised.

Review before enabling automatic publication. Use /commit status to preview or /commit local to commit without pushing. Require a verified remote and explicit current-branch destination before automatic pushes, including checks for broader push configuration.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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 (3)

Shadow Command Trigger

Medium
Category
Trigger Abuse
Confidence
70% confidence
Finding

Shadow Command Trigger: description clause '提交成功后自动推送当前分支到远端;正常情况全自动不询问,命中异常规则(不能或无须提交的文件等)才请用户确认。 TRIGGER(中文):"提交一下代码""帮我 commit""这次改动提交了吧""生成 commit message 并提交""/commit""按 fix 提交" "先看看变更别提交""改一下上次的 commit"。 DO NOT TRIGGER:用户只想看变更且未表达提交意图(直接执行 git status/diff 即可); 单独的 push、pull、rebas' conflicts with built-in command 'commit'

Content

No source excerpt is available for this finding.

The fallback commit-message format requires Chinese descriptions without user language opt-in.

Medium
Category
Output Handling
Confidence
99% confidence
Finding

When the project has no existing commit style, the workflow explicitly requires an English type and a Chinese description. This imposes a natural-language requirement without offering the user a language choice or documenting a locale-specific justification.

Content

No source excerpt is available for this finding.

The push command does not enforce the manifest's current-branch-only scope.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

The skill promises to push only the current branch, but a bare git push delegates destination and ref selection to Git configuration. Settings such as push.default=matching, configured push refspecs, or mirror behavior can cause additional branches or refs to be published; automatic execution without confirmation makes this scope mismatch more consequential.

Content

No source excerpt is available for this finding.