원클릭으로
pipeline-skills
pipeline-skills에는 johnrtipton에서 수집한 skills 14개가 있으며, 저장소 수준 직업 범위와 사이트 내 skill 상세 페이지를 제공합니다.
이 저장소의 skills
Execute pipeline stages by reading the state file and spawning agents. Reads .pipeline-state/<branch>.json, finds the next pending stage, spawns an agent to execute it, updates the state file, and repeats. Run /pipeline-next first to create the state file.
Shared pipeline stage procedures used by pipeline-feature, pipeline-bugfix, and pipeline-refactor skills. Not invoked directly — referenced by pipeline-specific skills as reusable building blocks for environment setup, testing, security scanning, documentation, commit/PR, code review, merge, and retrospective stages.
Drain open GitHub issues into the current milestone and process them all. Checks for open issues, adds them to the ROADMAP milestone, then runs pipeline-run --milestone --all to process everything.
Pick the next task (or group of related tasks) from ROADMAP.md and set up a pipeline state file. Reads the roadmap, filters by milestone/priority, groups related small features into batches, checks what's already done, and creates .pipeline-state/<branch>.json from the template. Run /pipeline-run after this to execute the stages.
Ship existing changes through quality gates to a merged PR. Starts from uncommitted/committed changes in your working tree and runs test → review → security → docs → commit → PR → code review → merge → retrospective. Use when you've been coding interactively and want to formalize and ship.
Fast iteration loop for initial-development work: branch → implement → build → push → live-verify → merge. No subagent reviews, no CHANGELOG, no roadmap coupling. Optimized for rapid PR cycles where the USER is the reviewer (manual verification via browser/CLI) and browser-extension reload/reconnect is an explicit handoff step. Use this INSTEAD OF /pipeline-ship when there's no CI, no formal review, and the goal is shipping many small PRs per day.
Initialize any repository so the entire pipeline-* family (next, run, dev, drain, retro, ship, strategy, roadmap-audit) works. Detects the repo's default branch, test/lint/build commands, ROADMAP format, and version scheme, then creates .pipeline-state/, .pipeline-templates/ (COPIED from the canonical pipeline-skill templates with the detected branch + commands substituted), gitignore entries, CHANGELOG/RETRO scaffolding, and a repo-root CLAUDE.md pipeline-config block — migrating a foreign-format ROADMAP to the canonical priority-matrix + milestone format the pipeline.py parser actually reads. Catches the failure where pipeline-next/run abort because no state template or conforming ROADMAP exists. Run once per repo before any other pipeline-* skill, or to repair a partially-set-up repo.
Run milestone retrospectives, maintain the action tracker, and reconcile deferred findings across RETRO.md, PR retros, and GitHub issues. Use after completing a milestone or when the user asks to review retros.
Verify every "not started" / "Not started" / non-✅ / non-~~strikethrough~~ ROADMAP.md entry against the actual codebase. Catches stale entries where a feature has shipped but the ROADMAP still lists it as pending. Run before cutting a release candidate, before starting a milestone, or after any large-scale consolidation.
Run a strategy session — the planning equivalent of /pipeline-run. Surveys the project state, brainstorms candidate features from pluggable sources, triages them, clusters into possible milestones, presents >= 2 distinct paths forward, recommends one, captures the chosen path back into ROADMAP and (if directional) a new ADR. State-file-as-program pattern: each stage is gated by a mandatory checklist on disk, same as /pipeline-run.
Automatically pull tasks from a ROADMAP.md file and process them through pipeline skills into a development branch. Parses milestone features, selects work by priority/milestone, and runs pipeline-feature or pipeline-bugfix for each task. Invoke with /pipeline-auto followed by optional filters.
Execute a Bug Fix pipeline: environment check, diagnosis, fix, test, regression check, documentation, commit & PR, code review, test verification, review verdict, merge, and retrospective. Invoke with /pipeline-bugfix followed by the bug report.
Execute a full Feature Implementation pipeline: environment check, planning, TDD implementation, test, self-review, security scan, documentation, commit & PR, code review, test verification, review verdict, merge, and retrospective. Invoke with /pipeline-feature followed by the task description.
Execute a Refactor pipeline: environment check, analysis, refactor execution, test, review, documentation, commit & PR, code review, test verification, review verdict, merge, and retrospective. Invoke with /pipeline-refactor followed by the refactor description.