The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): The skill provides numerous templates and executable examples for gh (GitHub CLI) and git commands. Specifically, the inclusion of gh pr create and gh pr merge --squash grants the agent the capability to perform state-changing operations on remote repositories.
[INDIRECT_PROMPT_INJECTION] (HIGH): The skill demonstrates a high-risk surface for indirect prompt injection when handling external content.
Ingestion points: The agent is expected to process external data including issue titles (e.g., Closes #123), PR descriptions, and code changes.
Boundary markers: None. The skill does not provide delimiters or instructions to ignore embedded instructions within the data it processes.
Capability inventory: Modifies repository state via gh pr create and gh pr merge (found in SKILL.md).
Sanitization: None. There is no guidance on escaping shell metacharacters or validating inputs before they are passed to the gh CLI as arguments (e.g., inside --title or --body). An attacker could inject malicious shell commands or override agent behavior by crafting a malicious issue or commit message.

github-pr