The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the gh CLI to interact with GitHub for tasks such as verifying repository details, checking user permissions, and creating issues. It also employs standard shell commands like mktemp, cat, and rm to manage temporary issue body files.
[INDIRECT_PROMPT_INJECTION]: The skill ingests issue template files from the project repository's .github/ISSUE_TEMPLATE/ directory. If these templates are maliciously crafted, they could contain instructions that redirect the agent's actions or cause it to leak information during the issue creation process.
Ingestion points: Template files are read from the .github/ISSUE_TEMPLATE/ directory within the target repository.
Boundary markers: There are no boundary markers or specific instructions for the agent to ignore embedded commands within the template files; instead, the skill explicitly tells the agent to treat these files as the "only source of truth."
Capability inventory: The skill possesses the ability to execute shell commands (gh) and read/write files in the /tmp directory.
Sanitization: The content extracted from the repository's templates is not sanitized or validated before being processed by the agent.

gh-create-issue