Sungblab

Initialize or bootstrap a repository with Devflow Native presets by calling devflow init and interpreting the result.

Load Devflow Native execution rules and repeated-mistake memory before running shell, browser, MCP, or test commands.

Run configured Devflow Native verification gates and record pass or fail evidence for a work item.

Inspect, install, verify, or repair native Devflow harness files for Codex, Claude Code, MCP, hooks, skills, gates, and review requirements.

Request, perform, or record Devflow Native review evidence before a task can be claimed done.

Inspect Devflow Native project truth, including repo state, work items, sessions, gates, review requirements, and latest handoff context.

Create, start, update, ready, block, unblock, rename, or list Devflow Native work items from local repo state.

Use when finishing a Devflow Native task from Codex; records evidence, checks documentation impact, respects Codex goal state, and asks the maintainer for commit/PR/continue direction.

Draft Korean blog retrospective posts. Use when the user wants a personal blog post about a project, learning process, failure, or changed judgment.

Draft Korean Facebook-style posts. Use when the user wants a personal, slightly longer public post for friends, communities, or project updates.

Draft short Korean Instagram Story text. Use for casual story captions, launch snippets, progress notes, or screenshot overlays.

Draft Korean or English LinkedIn-style professional posts. Use for project updates, lessons learned, launches, and career-facing reflections.

Draft Korean professor-facing messages. Use for emails, LMS notes, schedule requests, clarifications, and polite academic communication.

Draft Korean social thread posts. Use when the user asks for Threads/X-style serial posts, launch threads, or follow-up posts with per-post length limits.

Create a reusable writing SKILL.md for any recurring writing situation such as Threads posts, Facebook posts, LinkedIn posts, Instagram stories, professor messages, blog retrospectives, launch posts, or any user-defined format. Use when the user says they keep repeating writing rules or wants a new tone/channel/workflow skill.

{{description_yaml}}

Use before risky Claude Code edits or completion claims to catch agent boundary failures: wrong-scope edits, fallback before root cause, test-passing shortcuts, untrusted evidence leakage, oversized plans, or missing verification evidence.

Use before risky Codex edits or completion claims to catch agent boundary failures: wrong-scope edits, fallback before root cause, test-passing shortcuts, untrusted evidence leakage, oversized plans, or missing verification evidence.

Use when coding-agent work risks boundary failures: user intent or copy role confusion, fallback before root cause, named tool changes, test-passing shortcuts, scope creep, untrusted context, or completion without evidence.