The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a shell command template for creating GitHub issues: gh issue create --repo n8n-io/n8n --title "<title>" --body "$(cat <<'EOF' ... EOF)". The use of double quotes for the title parameter derived from user input allows for potential shell command injection if the input contains shell metacharacters such as backticks or $().\n- [PROMPT_INJECTION]: The skill processes untrusted user data to generate structured content for external platforms, representing an indirect prompt injection risk.\n
Ingestion points: User input provided via the $ARGUMENTS parameter and data retrieved from mcp__notion__notion-search.\n
Boundary markers: The skill uses a bash heredoc (<<'EOF') for the issue body, which helps isolate content from shell execution, but no similar markers or delimiters are defined for the issue title or the content's effect on downstream systems.\n
Capability inventory: The skill can execute shell commands via the gh CLI and perform write operations to external services using mcp__linear-server__save_issue.\n
Sanitization: While the skill requires a human-in-the-loop confirmation before final execution, it lacks specific instructions for the agent to sanitize or validate the content of user-provided strings for shell safety or prompt injection prevention.

create-issue