Vultron

End-to-end PR review pipeline. Detects the PR on the current branch, then runs pr-triage → pr-execute → pr-verify in sequence. Supports resume: if artifacts already exist from a prior run, skips forward to the right phase. No user prompts except the one deferred-ask pause in pr-execute. Precondition: an open PR must exist for the current branch (or provide an explicit PR number).

Snapshot the current state of all work queues and recommend the next skill to run. Reads BUILD_LEARNINGS.md, GitHub Issues (by type: Idea, Bug, Concern), unscheduled issues, PRIORITIES.md, and the git log to assess staleness. Produces a concise status report, then asks the user whether to invoke the recommended skill. Use when resuming work after any interruption, or when unsure what to do next in the development cycle.

Execution phase of the PR review pipeline. Reads .claude/pr-{number}-triage.json, applies all FAIL/IMPROVE fixes inline, remediates CI failures, files GitHub issues for out-of-scope findings, resolves review thread comments, and writes .claude/pr-{number}-execute.json. Use after /pr-triage, or as the second step of /pr-ship.

Pure-discovery phase of the PR review pipeline. Runs all 11 inspection phases (orient through CI status), produces a machine-readable finding list at .claude/pr-{number}-triage.json, and posts a human-readable PR comment. Makes NO code changes. Use as the first step of /pr-ship, or standalone when you want a full finding inventory before deciding whether to execute fixes.

Verification phase of the PR review pipeline. Reads .claude/pr-{number}-execute.json, spot-checks each claimed fix against HEAD (not just the commit diff), validates CI and test suite, and posts a per-finding verdict comment on the PR. Cleans up both artifact files as its final step. Makes NO code changes. Use after /pr-execute, or as the third step of /pr-ship.

DEPRECATED — use /pr-ship (full pipeline) or /pr-execute (fix phase only) instead. Kept as reference for testing logic and comment resolution strategies absorbed into pr-execute/REFERENCE.md; do not invoke directly for new work.

DEPRECATED — use /pr-ship (full pipeline) or /pr-triage (discovery only) instead. Kept as canonical prose reference for the 11 inspection phases; do not invoke directly for new work.

Fix a bug using test-first development. Gates implementation on confirmed shared understanding with the user — no code is written until both the agent and the user agree on what bug is being fixed and why. Use when the user asks to fix a bug.

Completes the highest-priority pending implementation task: loads project context, selects the next task from GitHub Issues, implements it, validates, opens a PR, and updates plan history. Use when the user asks to continue planned implementation work or turn the next prioritized item into a completed changeset.

Promote lessons learned from the build process into durable specifications and design notes. First refreshes docs/reference/codebase/ via acquire-codebase-knowledge, then reads plan/incoming/learnings/ and queries GitHub for open type:Concern issues (both as input queues), analyzes gaps, interviews the user with grill-me to align on scope, then writes to specs/, notes/, and AGENTS.md, opens a docs-only PR with the specs-notes label, and archives processed entries. Use when build execution has produced insights that should be reflected in specs or notes. For external ideas (GitHub Idea-type issues), use `plan-issue` instead.

Branch-freshening PR submission. Cherry-picks the task branch onto a fresh origin/main base immediately before pushing, validates (linters for docs PRs; full suite for implementation PRs), and opens a PR. Called by build, plan-issue, bugfix, and learn. Can also be invoked conversationally ("pr that for me"). Returns the PR URL.

Load the always-required baseline context before any implementation, planning, or documentation work. Reads the glossary, loads all specs, reads AGENTS.md, the completeness doctrine, notes/README.md, and docs/adr/index.md, reads plan/incoming/learnings/, and queries Project #24 for Schedule=Now items. Run this at the start of every workflow skill before selecting or reading a specific issue. Replaces study-project-docs Phase A.

Archive a single completed work item to plan/history/ using the append-history tool, then lint, stage, commit, and push the resulting files. Always call this skill AFTER a PR has been created (so the PR URL can be included in the entry body). Caller constructs the entry body and passes it as a heredoc. Handles one entry at a time; callers that need to archive multiple items (e.g., learn) invoke this skill once per item in a loop. Use whenever any other skill calls `uv run append-history`.

Convert a single open GitHub Idea, Concern, or Epic issue into a concrete implementation plan. Deepens context from the issue first, then runs a grill-me interview to understand scope, creates implementation issues, optionally updates specs/notes, and closes or annotates the source issue appropriately. Auto-detects type from the GitHub issue type. Use when the user references an Idea, Concern, or Epic issue number, or says "plan this idea/concern/epic".

Captures all work-in-progress context when a bugfix session is interrupted and posts a structured handoff comment to the GitHub issue, then pushes the branch. Use when a bugfix agent must stop before fully resolving the bug — session ending, agent looping, developer leaving, or any other interruption that requires handing off to a future worker. Immediately halts all resolution work and shifts to documentation.

Create or update a GitHub Issue with structured relationships (parent/child sub-issues, blocking/blocked-by) using the GraphQL API. Idempotent for all relationship wiring. Strips legacy body-text markers ("Blocked by #N", "Blocks #N", "Parent: #N") when structured relationships are applied. Call this skill wherever any other skill creates or modifies issues that carry relationships.

Create a GitHub Epic issue for a priority group and wire existing leaf issues as its sub-issues. Uses the GitHub Epic issue type (via GraphQL) and the GraphQL addSubIssue mutation. Invoke this skill whenever a thematic group needs an Epic created or updated.

Batch-process docs/reference/codebase/CONCERNS.md into GitHub Concern-type issues. Optionally runs a focused acquire-codebase-knowledge scan first. Deduplicates against existing open Concern issues — updating the body and appending a refresh comment on matches, creating new issues otherwise. Handles the [ASK USER] Questions section interactively. Adds all new issues to Project #24 with Schedule=Someday. Does not write to specs/, notes/, or open a PR. Use when you want to turn a codebase scan into a set of tracked GitHub issues.

Thin git stage-and-commit wrapper. Stages the specified files and commits with a clear message and the required Co-authored-by trailer. Does not perform finalization (history updates, plan cleanup) — the caller owns that. Use after all finalization and validation steps are complete.

Perform a gap analysis between current specs/notes and the codebase, then create GitHub Issues for any untracked gaps and add them to Project #24. Observations and open questions go directly to notes/ files. Use after learn or `plan-issue` has updated specs/notes, and before running build.

Create or update a single GitHub Idea or Concern issue from freeform input, with type inference, duplicate checks, and optional Epic parent wiring. Use when a developer wants to quickly capture a new planning item as type:Idea or type:Concern.

Load task-specific context after the target issue is known. The caller passes focus hints (e.g., "wire layer", "BT integration", "embargo lifecycle") and this skill reads the relevant notes files, ADRs, and codebase reference files. Run after orient-agent, once the issue to be worked has been selected and read. Replaces study-project-docs Phase B.

Condense AGENTS.md files that have grown too large by migrating content to per-directory AGENTS.md files and notes/ references. Use when root AGENTS.md exceeds ~400 lines, or when asked to clean up, condense, or restructure AGENTS.md.

Audit and report on priority status against the GitHub Project #24 board, open issues, and PRs — tracking progress, coverage gaps, stale items, and readiness by Schedule tier (Now/Next/Later/Someday). Use when reviewing where the project stands relative to the board, or before running review-priorities.

Audit and update Project #24 ("Vultron Planning") in one workflow. Runs a comprehensive status check against the board and open issues/PRs, then offers interactive options to move items between Schedule tiers (Now/Next/ Later/Someday), promote Triage items, archive completed Epics, and wire sub-issue relationships. Use when you want to review where the project stands and make any necessary board updates.

Interactively update GitHub Project #24 ("Vultron Planning") by moving items between Schedule tiers (Now/Next/Later/Someday) via the GitHub Projects API. Use when you want to schedule issues, promote triage items, or rebalance tiers based on findings from check-priority-status.

Generate a project report summarizing project progress for a given time period (month, quarter, or year-to-date). Gathers data from plan/history, git log, and GitHub PRs, then produces a capability-focused narrative with an executive summary and section headings. Use when the user asks for a project report, sponsor update, or monthly/quarterly summary — e.g. "/project-report May 2026" or "/project-report Q1 2026".

Use this skill when the user explicitly asks to map, document, or onboard into an existing codebase. Trigger for prompts like "map this codebase", "document this architecture", "onboard me to this repo", or "create codebase docs". Do not trigger for routine feature implementation, bug fixes, or narrow code edits unless the user asks for repository-level discovery.

Interview the user relentlessly about a plan or design until reaching shared understanding, resolving each branch of the decision tree. Use when user wants to stress-test a plan, get grilled on their design, or mentions "grill me".