Skills
skills/maxmurr/agents-skills/close-issue/Gen Agent Trust Hub

close-issue

Warn

Audited by Gen Agent Trust Hub on Mar 26, 2026

Risk Level: MEDIUMCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run the shell command gh pr view <pr-url> using a user-supplied parameter without quoting or sanitization. This allows for arbitrary command execution if a malicious string (e.g., ; rm -rf /) is provided. This also constitutes an indirect prompt injection surface. 1. Ingestion points: SKILL.md via the <pr-url> placeholder. 2. Boundary markers: Absent for the command invocation. 3. Capability inventory: Execution of shell commands via the gh tool. 4. Sanitization: No quoting or validation is performed on the user input before its use in the shell.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 26, 2026, 06:36 AM