pc-style

Keep one agent task progressing linearly until an objective completion condition passes, with fresh state checks and hard resource limits. Use when the user has one long-running goal that does not need parallel PR orchestration or a changing workflow graph.

Watch one existing GitHub pull request, validate and address review comments, repair scoped CI failures or clear conflicts, request fresh review after each changed head, and stop when merge-ready or blocked. Use when the user asks to babysit, monitor, fix, or carry a PR through review.

Orchestrate a multi-stage engineering project through dynamically grouped or stacked Codex tasks, branches, pull requests, independent reviews, remediation cycles, and merge. Use when one goal needs multiple PRs or when the user wants a project carried autonomously from plan to merged changes.

Maintain a repository on a recurring heartbeat by inspecting issues, pull requests, CI, dependencies, and repository-local priorities, then directing bounded work to fresh Codex tasks. Use when the user asks Codex to maintain a repo, run recurring engineering triage, or wake periodically and dispatch useful work.

Design and run a loop around any repeatable agent workflow. Use when the user invokes `$loop`, asks Codex to keep watching or working, wants agents to prompt other agents, or needs a dynamic workflow whose shape should be inferred from the goal rather than selected up front.

Monitor a changing source or condition on a cadence, persist the last meaningful observation, and notify or act only when a user-defined change occurs. Use for watching pull requests, issues, releases, prices, availability, feeds, pages, dashboards, or other external state.

Routes pcstyle's recurring cross-project tasks to compact workflow references. Use for fuzzy multi-step work involving design directions, browser visual verification, audits, extraction, benchmark fixtures, PDF/document production, API discovery, client-source ingest, or continuation from prior threads.

Keep a PR merge-ready by triaging comments, resolving clear conflicts, and fixing CI in a loop.

Builds polished, user-friendly shell scripts and install.sh flows. Use when creating or editing Bash, Zsh, POSIX sh, setup scripts, installer scripts, CLI workflows, or interactive terminal automation; prefers Charmbracelet gum unless there is a concrete reason not to.

Create or resume from a handoff message for continuing work in a new AI session. Use when the user wants to save context for later, create a handoff document, continue from an existing HANDOFF.md, or prepare work for continuation in a fresh session. Captures relevant files, context, and goals so the next session can start immediately without rediscovering the codebase.

Manages repository-scoped secrets and environment-variable memory with Charmbracelet skate. Use when discovering env vars, API keys, Vercel project variables, generated service credentials, or when asked to enable/disable skate-backed secret memory for a repo.

Create or update AgentSkills. Use when designing, structuring, or publishing skills with references and assets. Triggers on requests like "create a skill", "build a skill for X", "publish this skill", or when the user wants to design, structure, or publish AgentSkills.

Turn rough feedback about the Warp app into a filed GitHub issue or duplicate-issue response for `warpdotdev/warp`. Use when the user shares a Warp bug report, regression, confusing UX note, feature gap, or short complaint and wants it clarified into reproduction steps, expected vs actual behavior, concrete impact, and grounded source references from local Warp repos when available.

Guides testing Warp UI features and changes using the computer use tool. Use this skill only when the computer_use tool is available to the agent. Covers launching Warp and verifying UI behavior.

Invoke this automatically after completing any user-facing client change, ONLY in non-sandboxed environments and local environments. Verifies changes by spawning a cloud agent with computer use to test Warp. Triggers a cloud agent that runs the test-warp-ui skill.

Address review comments on the current branch's GitHub PR via gh CLI.

Run automated code review using cubic review with amp review fallback. Use when the user wants to review code changes, check PR quality, analyze code for issues, or run automated code review tools.

Generate AGENTS.md file for a codebase to help AI agents understand the project structure, build commands, testing, and conventions. Use when the user asks to create an AGENTS.md, initialize agent documentation, or analyze a codebase for agent context. Also use when setting up a new repository for AI-assisted development.

Route npm and VS Code extension installs through Supply Chain Guard (scguard). Use when adding dependencies, running package managers, or installing extensions. Never call bun/npm/pnpm/yarn install or add directly when this project uses scguard.

Reads public X (Twitter) statuses and conversations, profiles, search results, followers, and following through x.pcstyle.dev. Use when an agent needs X content as compact or full Markdown or JSON without cloning a repository.

Splits work across multiple Amp sub-agents or Orbs and syncs their changes back locally. Use when the user asks to parallelize, use sub-agents, multiple agents, Orbs, fleet, fan out work, run phases concurrently, or combine parallel work with amp sync.

Can be used to kick off local "run loops" on the RWX CI platform, or when understanding, creating or modifying an RWX CI/CD config or any other file in a .rwx directory. TRIGGER when: the user asks about RWX; wants to create or modify an RWX config; says things like "CI failed", "CI is red", or "tests are failing in CI"; asks "what's failing on my PR / branch / commit?"; wants to debug test failures, check a run status, fetch logs/artifacts, or start a run loop.

Access the public Librarian research service for source-backed answers about GitHub repositories and public web pages. Use this skill whenever the user asks you to research a GitHub repo, inspect public source files, search commit history, compare repository diffs, read public docs/pages, or answer questions that need concise citations from repo/web evidence. This is especially useful when you do not have direct network access, when you want a dedicated research subagent, or when the user mentions Librarian.

Use when the user says "cold <skill-name or query>" (e.g. "cold to-prd", "cold something that writes prds"), "Search your skills", or asks to find/use a cold, hidden, or non-autoloaded skill. Searches ~/.agents/skills-cold and manually loads matching skill instructions on demand.

Delegate narrow implementation slices from a parent smart thread into locked-down rush child runs. Use when the parent can define an explicit child contract with task, context paths, write paths, create paths, acceptance criteria, allowed commands, and allowed skills, and when the parent will retain final review and integration. Do not use when scope is broad, cross-cutting, ambiguous, decision-heavy, or requires broad tool, MCP, or network access.

Orchestrate parallel sub-tasks by spawning non-interactive instances of your own CLI as subagents. Use when you need to parallelize work across multiple files, run independent investigations simultaneously, or delegate heavy multi-step tasks. Works with ANY AI coding CLI agent (Amp, Claude Code, Codex, Cursor, OpenCode, aider, Cline, Roo, goose, Windsurf, Copilot CLI, pi, etc.). Triggers on "run in parallel", "subagent", "delegate", "fan out", "concurrent tasks", or any complex task that benefits from parallel execution.