blueprint

Push the current branch to remote with branch safety checks. Triggers on "/bp-push", "/push", "push this", "push to remote", "push branch", or any request to push code to the remote repository. Also triggers on "push my code", "push the branch", or "send it to remote". Does NOT trigger on "push my changes" — use /bp-ship if uncommitted changes exist. Blocks pushes to main/master — only staging and feature branches are allowed.

Fully autonomous SDLC pipeline — zero user intervention. Opus 4.6 runs the entire workflow from plan through PR, handles review feedback loops, and stops only when the PR is ready for human merge. Use this skill whenever the user says "/flow-auto", "auto flow", "autonomous flow", "just do everything", "full auto", "hands off", "run it all", or any request to run the complete pipeline without checkpoints or user decisions. Also triggers on "no intervention", "auto pipeline", "unattended flow", "fire and forget flow", "do everything and leave me a PR", or "I'm going to sleep just build it". This is the zero-touch version of /flow — same pipeline, no pauses. IMPORTANT: never use AskUserQuestion in this skill — all decisions are made autonomously.

Auto-chain the full SDLC workflow with checkpoints and pauses for review. Use this skill whenever the user says "/flow", "run the full workflow", "auto chain", "start to finish", "full pipeline", "plan and execute", or any request to run the complete plan-to-finish pipeline automatically. Also triggers on "chain skills", "workflow pipeline", "run everything", "take this from plan to PR", "full dev cycle", or "run plan through finish". NOT for autonomous/zero-intervention flows (use /flow-auto instead). Orchestrates: plan → review → approve → check → PR → finish.

Emergency hotfix deployment: commit, push to staging branch, create PR to main, and merge. Use this skill when the user says "/hotfix", "emergency push", "hotfix to production", "push to staging and merge to main", "urgent deploy", "fast push to main", "deploy hotfix", or any request for an emergency/urgent deployment bypassing the normal /plan → /pr → /finish workflow. This is the fast lane: commit → push → PR → merge.

Trigger @claude PR review on a Pull Request via GitHub Actions. Use this skill whenever the user says "/review", "review the PR", "code review", "trigger review", or any request to get an automated code review on a PR. Also triggers on "claude review", "run review", or "check the PR". Posts a comment mentioning @claude which triggers the GitHub Action.

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Fetch PR review comments, categorize feedback, implement fixes, and push. Use this skill whenever the user says "/address-pr", "address review", "fix review comments", "handle PR feedback", or any request to implement fixes based on PR review feedback. Also triggers on "fix the review", "address feedback", "implement review changes", "implement the requested changes", or "handle review comments". Fetches real comments via blueprint pr-review.

Capture, manage, and promote pre-planning ideas before they become full plans. Use this skill whenever the user says "/backlog", "add to backlog", "backlog idea", "capture this idea", "save this for later", "I had an idea", or any request to manage the idea backlog. Also triggers on "list ideas", "promote idea", "backlog list", "drop idea", "mark done", "what's in the backlog", "show me the ideas", "pending ideas", or any request to add/view/promote/done/drop items in blueprint/backlog/. The backlog is the pre-planning stage — ideas live here until promoted to /plan. Even if the user doesn't explicitly say "backlog", trigger when they describe wanting to capture a feature idea, task, or improvement for later without implementing it now. This skill is the gateway to the SDLC — nothing gets planned or built without first passing through the backlog.

Execute multiple plans sequentially in a single session — loops flow-auto for plans N through M with auto context compaction, merge chain automation, and crash recovery. Use this skill whenever the user says "/batch-flow", "batch flow", "run plans 2 to 6", "execute all plans", "sequential plans", or any request to run multiple plans in one session. Also triggers on "multi-plan", "batch execute", "run remaining plans", "plans N through M", "batch pipeline", "run plans 3 4 and 5", "I have N plans to execute", or "execute plan N through plan M". This is the multi-plan wrapper around /flow-auto. IMPORTANT: never use AskUserQuestion in this skill — all decisions are made autonomously.

Create a new feature branch following type/kebab-case conventions. Triggers on "/bp-branch", "/branch", "create branch", "new branch", "make a branch", or any request to create a git branch. Also triggers on "checkout new branch", "start a branch", "feature branch", "I need a new branch", or "branch for [something]". Uses type/description format: feat/feature-name, fix/bug-name, etc.

Commit current changes following emoji+type conventions with code formatting. Triggers on "/bp-commit", "/commit", "commit this", "commit changes", "save my changes", or any request to create a git commit. Also triggers on "stage and commit", "commit with message", "make a commit", "let's commit", or "save this progress". Do NOT trigger when the user also mentions pushing — route to /bp-ship instead. Runs code formatting on changed files before committing.

Scan project and generate/audit CLAUDE.md files with stack auto-detection and Context7 docs. Triggers on "/bp-context", "/context", "generate context", "audit CLAUDE.md", "scan project", or any request to create or update CLAUDE.md documentation files. Also triggers on "brownfield", "onboard project", "onboard this project", "onboard existing project", "scan existing project", "update context docs", "context scan", "check CLAUDE.md", "stale documentation", or "refresh project docs". Runs parallel workers per directory. Also audits README.md for staleness.

Commit and push changes in one go with branch safety checks. Triggers on "/bp-ship", "/ship", "ship it", "commit and push", "push my changes", or any request to commit and push in a single step. Also triggers on "ship this", "send it", "deploy changes", "save and push", or "let's ship". This is /bp-commit + /bp-push combined with a branch safety gate.

Show quick repo status — branch, base, plan, PR info in one shot. Triggers on "/bp-status", "/status", "show status", "repo status", "what branch am I on", "current status", "where am I", "project status", "what's the current state", "which branch", "is there a PR open", "what plan am I on". Uses blueprint CLI with git fallback.

TDD variant of plan-review — restructures the plan to follow Red/Green/Refactor cycle. Triggers on "/bp-tdd-review", "/tdd-review", "tdd plan", "test-driven review", "red green refactor", or any request to prepare a plan for TDD execution. Also triggers on "tdd mode", "write tests first", "test-first approach", "tdd planning", "prepare plan for TDD", or "write tests before code". Injects Phase 0 (all tests written first, failing), GREEN implementation phases, and final Coverage & Refactor phase targeting 100% coverage.

Create tests following project conventions with proper factories and real implementations. Triggers on "/bp-test", "/test", "write tests", "create tests", "add tests for", or any request to create new test files. Also triggers on "test this", "write a test for", "feature test", "unit test", or any request to generate test code. Uses the project's test runner and conventions detected from configuration.

Clean up worktree after PR merge — only needed for /plan-wt flows. Use this skill whenever the user says "/complete", "clean up worktree", "remove worktree", "delete the worktree", "worktree cleanup", or any request to clean up after a worktree-based feature is merged. For standard /plan flows, /finish handles everything — /complete is a no-op. Detects environment automatically and skips if nothing to clean up.

Mark feature complete, merge PR, move plan to upstream, and handle staging→main flow. Use this skill whenever the user says "/finish", "finish this", "merge and finish", "wrap up", "close this out", or any request to complete a feature and merge the PR. Also triggers on "merge PR", "finish the feature", "we're done", "mark as complete", "the PR is approved let's merge", or "feature is complete". Run AFTER PR is approved. Handles plan move to upstream and main merge.

Fully autonomous SDLC pipeline in an isolated git worktree — zero user intervention, zero git conflicts with other terminals. Creates a lightweight worktree (no Herd URL, no VS Code, no SSL), runs the full flow-auto pipeline inside it, and stops when the PR is ready for merge. Use this skill whenever the user says "/flow-auto-wt", "auto flow worktree", "parallel flow-auto", "flow-auto in worktree", or any request to run the autonomous pipeline in an isolated worktree. Also triggers on "isolated flow-auto", "worktree auto", "parallel auto", "run flow-auto in isolation", "don't touch my branch", "run in parallel", "parallel pipeline", or when running multiple flow-auto instances on the same repo. IMPORTANT: never use AskUserQuestion in this skill — all decisions are made autonomously.

Initialize a new project with GitHub Actions, git hooks, blueprint workspace, CLAUDE.md, and branches. Use this skill whenever the user says "/start", "initialize project", "setup project", "start new project", or any request to set up the standard development workflow in a repo. Also triggers on "project init", "bootstrap project", "setup hooks", or "initialize repo". Sets up everything: claude-pr-reviewer.yml + tests.yml actions, commit hooks, blueprint/ workspace, CLAUDE.md, configurable staging branch. Updates are handled by `blueprint update` — no separate sync skills needed.

Execute a reviewed plan by delegating work to subagents and coordinated teams. Reads the Execution Strategy from the plan, dispatches workers per round, tracks task completion with [x] marks and timestamps, commits after each phase, and manages context for long executions. Use this skill whenever the user says "/plan-approved", "execute the plan", "start implementation", "run the plan", or any request to begin coding based on an existing reviewed plan. Also triggers on "let's build it", "go ahead and implement", or "start coding". Does NOT trigger on: "resume", "continue the plan", "pick up where we left off", "where did we stop" — use /resume for those. ALWAYS run after /plan-review. With 1M context, if context usage is under 30% after plan-review, continue directly — the coordinator only orchestrates (delegates to subagents), so it needs minimal context. Only clear when context is already above 50% after plan-review.

Audit implementation against the plan — compare planned vs actual changes, detect orphaned test references, fix checkbox mismatches, and sync frontmatter counts. Use this skill whenever the user says "/plan-check", "check the plan", "audit the plan", "verify implementation", "compare plan vs code", or any request to validate that what was built matches what was planned. Also triggers on "orphaned tests", "plan audit", "check task marks", "did we implement everything", or "sync frontmatter counts". ALWAYS run after /plan-approved and before /pr — this is the quality gate.

Review, validate, and prepare a plan for optimal execution by subagents or coordinated teams. And also Assign Execution Plan — this skill both validates the plan AND determines the execution strategy (parallel subagents, teams, mixed dispatch), marks task complexity ([H]/[S]/[O]), and commits the reviewed plan with its assigned execution mode. Use this skill whenever the user says "/plan-review", "review the plan", "prepare plan for execution", "optimize the plan", "is the plan ready", "plan readiness", or any request to validate and prepare an existing plan before implementation. Also triggers on "mark complexity", "execution strategy", "plan delegation", or "prepare for plan-approved". ALWAYS use this skill after /plan and before /plan-approved — it's the pre-flight check that makes execution fast. Do NOT use for post-implementation audits — use /plan-check for "check the plan" or "audit the plan". With 1M context, small/medium plans (≤15 tasks) can flow directly into /plan-approved without clearing context.

Create structured implementation plans with git branches and memory search. Use this skill whenever the user wants to plan a feature, fix, refactor, or any development task. Triggers on: "/plan", "plan this", "create a plan for", "let's plan", "I need to plan", "break this into phases", "how should we approach", "let's think about how to build", or any request that involves planning, phasing, or strategizing before coding. For worktree-based plans (isolated directory), use /plan-wt instead. ALWAYS use this skill before starting implementation — planning first, coding second.

Create a Pull Request with plan context, GitHub Issues integration, and proper base branch detection. Use this skill whenever the user says "/pr", "create a PR", "open a pull request", "make a PR", "create pull request", or any request to create a PR for the current branch. Also triggers on "open PR", "submit PR", "PR for this branch", "push and create PR", "I'm ready for a PR", or "let's open a pull request for this". ALWAYS runs /ship first and detects the correct base branch (feature→{staging_branch}→main flow).

Quick task execution without a full plan — just do it and commit. Use this skill whenever the user says "/quick", "quick fix", "just do it", "small fix", or any request for a small change that doesn't need planning. Also triggers on "quick task", "small change", "tweak this", "simple fix", "rename this", "fix this typo", "update the config", "one-liner", "trivial change", "swap this", "change this value", "toggle this", or any 1-3 file change that doesn't warrant a full /plan workflow. Supports optional GitHub Issues integration for issue tracking.

Resume a plan after context limit, crash, or session break. Detects completed phases, summarizes progress, and continues execution from where it stopped. Use this skill whenever the user says "/resume", "continue the plan", "pick up where we left off", "resume plan", "context ran out", "ran out of context", "continue from where I stopped", or any request to resume an interrupted plan. Also triggers on "what was I working on", "plan status", "re-enter plan", "where did we stop", or returning to an in-progress plan after a break.

2 hooks across 2 events