csepulv
- 12 skills
- 0 followers
- 10 hours ago last updated
- ▌ Michi Loop · csepulv bundleSet up and launch a disciplined lights-off loop — gate readiness, define the goal/verifier/stop contract, hand off to a loop mechanism, review at halt. For convergent work with a verifier you can trust. The convergent + autonomous counterpart to michi-expedition.
- ▌ Michi Debrief · csepulv bundleStructured post-session review — assess what was delivered, review decisions, capture learnings, audit what got invalidated, refresh the root docs, and calibrate trust for the next session.
- ▌
- ▌ Michi Pr Prep · csepulv bundlePrepare a PR review guide — a companion document that helps reviewers understand what they're looking at, why decisions were made, and where to focus attention. Two output shapes (TLDR + Details by default, or TLDR-only via `/michi-pr-prep tldr`); two context modes (in-session, from-diff).
- ▌ Michi Session · csepulv bundleExecute a michi implementation session — implement a milestone with mandatory verification, decision logging, code review, and commit discipline. This is a rigid skill — follow exactly.
- ▌ Michi Planning · csepulv bundlePrepare a project milestone for michi execution — write the plan, define acceptance criteria, set up verification, and scope the work for autonomous agent implementation.
- ▌ Michi Workshop · csepulv bundleLightweight Michi discipline for everyday work — bug fixes, small features, quick explorations. Same rigor, less ceremony. The EDC of software development.
- ▌ Michi Bootstrap · csepulv bundleOnboard a project to the Michi documentation structure — survey what exists, recommend what to build (S/M/L/XL effort tiers), and walk the user through setup interactively.
- ▌ Michi Docs Site · csepulv bundleScaffold an internal Astro + Starlight docs browser for any project, or generate a PDF build recipe for an existing Starlight site. Workshop-style skill.
- ▌ Michi Expedition · csepulv bundleRun a Michi expedition — open-ended, Entrusted, spiral exploration where the end isn't known (data exploration, external research, experiments, prototypes) with a feedback flywheel. The learning-mode counterpart to planning/session/debrief. Modes: charter (open/revise a mission), campaign (set the mandate), run (the spiral), review (surface + promote), exhibit (the review surface). Use when the goal is to discover, not to ship a known deliverable.
- ▌ Michi Sustainability · csepulv bundleAssess and maintain system health — code quality, test quality, architectural alignment, and knowledge capture. Scales from a brief within-milestone check to a deep between-epic review.
- ▌ Michi Scenario Test Builder · csepulv bundleGenerate scenario-based test plans and test cases using Cem Kaner's scenario testing methodology. Use this skill when the user wants to discover what to test, create test scenarios, build test plans, define acceptance criteria, generate verification scenarios for autonomous agent work, or explore how a system might fail or be misused. Also use when the user mentions "scenario testing", "test plan", "acceptance criteria", "what should I test", "verification scenarios", "holdout tests", or asks how to validate that a feature works end-to-end. This is NOT for writing test code — it produces test plans, scenarios, and criteria that inform test implementation.