The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several shell commands to gather environment data and perform its primary function, including uname -a for OS information, claude --version or codex --version for agent versions, and gh issue create for submitting the report.
[PROMPT_INJECTION]: The skill processes untrusted data from user responses and interpolates it into the bug report and shell command arguments, which represents a potential surface for indirect prompt injection or command injection if the execution platform does not properly escape arguments.
Ingestion points: User responses to bug-related questions collected via AskUserQuestion or similar tools (SKILL.md).
Boundary markers: No explicit delimiters are used to separate user-provided content from the report template.
Capability inventory: Shell execution via gh, uname, and CLI version commands.
Sanitization: No explicit instructions for sanitizing or escaping user-provided text before interpolation into CLI arguments.

report-bug