Back to skill

Security audit

subclip-github

Security checks for vulnerabilities and agentic risk

Overview

The scanners rate this skill benign with high confidence: they report one low-probability concern about autonomous decisions, while the reviewed content describes confirmation before repository changes.

The scanners report complete coverage of the single listed Markdown component and no AIG findings. The verdict is benign, not risk-free; consider whether the reported GitHub access and confirmation process suit your needs before installing.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 31)May include surrounding context.

md
l_request` first and read its `commits` plus `exactDiff`. Use the connected read bridge afterward when private access, branch history, check runs, reviews, or additional commit metadata is needed.
- **Changelog in a date range:** combine releases/tags with commits or merged pull requests in the same range. Prefer release notes when present, group changes by feature/fix/maintenance only when evidence supports it, and link or identify the underlying releases, PRs, or commits returned by GitHub.

When a repository name is unique in the connected account, select it without asking for a redundant `owner/repo`. If multiple repositories match, show the short candidate list and ask which one.

Reads may run directly. Before creating or changing issues, pull requests, releases, files, branches, comments, stars, or repository settings, show the exact GitHub account, repository, action, and submitted fields and obtain confirmation unless the current message already authorizes those exact details.