The Agent Skills Directory

[SAFE]: The skill uses established tools, specifically the GitHub CLI (gh) and the Linear MCP, for its core functionality of pull request and issue management.
[COMMAND_EXECUTION]: The skill executes gh pr create to automate pull request drafting and submission, which is a legitimate and safe use of the tool.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes git history and branch names to generate PR content. Ingestion points: git branch names, commit messages, and diffs. Boundary markers: none. Capability inventory: GitHub CLI (network/API) and Linear MCP (network/API). Sanitization: none. This behavior is standard for this type of utility and is not considered a threat.

open-pr