orperelman123
UserPermission-safe prompt refinement and bounded graph orchestration for Codex, Claude Code, Cursor, and GitHub Copilot.
Categories
Indexed Skills (2)
engineer-graph
Plan, validate, and run bounded autonomous engineering graphs. Use for broad work that benefits from decomposition, parallel investigation, independent verification, synthesis, implementation, or a human gate before external or destructive actions.
improve-prompt
Convert user requests into intent-preserving, permission-safe execution briefs with useful context, requirements, constraints, acceptance criteria, and verification. Use before substantive implementation, investigation, research, or external actions; when the user asks to improve, rewrite, clarify, or evaluate a prompt; or when a request is ambiguous enough to benefit from a structured brief.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.