skills

Almanac reference for dates, holidays, time zones, business days, and astronomical events. Use when checking if a date is a business day, calculating date differences, converting time zones, looking up holidays, or checking for astronomical events like solstices, equinoxes, and moon phases. Also for seasonal calculations, workday schedules, and calendar arithmetic.

Toyota Production System (TPS) and Toyota Way for software engineering — lean thinking, waste elimination, root cause analysis, and continuous improvement. Use when identifying waste (muda/muri/mura), running blameless postmortems, doing 5 Whys, writing A3 problem statements, or connecting lean philosophy to TRIZ contradiction resolution.

Local multilingual translation via Ollama — 119 languages, no API key, no network, no data leaves the machine. Use when translating to/from Japanese, Chinese, Korean, French, Spanish, German, Portuguese, Russian, Arabic, or any other language. Also for localization, i18n text, and roundtrip validation. Qwen3:32b on Apple Silicon — faster and cheaper than cloud APIs.

TRIZ contradiction matrix oracle for software and systems design. Use when improving one quality worsens another — speed vs correctness, isolation vs sharing, autonomy vs control, simplicity vs completeness. Trigger phrases: "tradeoff", "tension between", "can't have both", "worsens when", "at the cost of", "faster but less reliable", "more isolated but harder to share". Not for bugs, merge conflicts, or tool choices. Prerequisite: triz binary must be installed (run `triz doctor` to verify).

Deploy and manage software on Linux/BSD fleet hosts (Pis, NAS, workstations) via Ansible. Use when user mentions deploy, ansible, fleet, hosts, inventory, or deployment.

Use when operating, debugging, or deploying the elixir-gastown Elixir/BEAM fleet — cluster federation, release builds, worker deploys, seccomp fixes, sname/name distribution issues, rate limiter, or any gastown service on Pi workers or coordinator nodes.

Use when communicating with fleet agents, checking agent status, or sending commands via IRC. The fleet runs TLS IRC on 10.0.3.203; agents listen in

Use when operating the fleet of Raspberry Pis, NAS hosts, and workstations via irclaw IRC bots, Ansible, Ollama (kimi-k2.5:cloud), or long-running deploys. Covers IRC DM protocol, bot security policy, zellij for long tasks, OS upgrades, and gastown/polecat deployments.

Use when spawning parallel research agents across the irclaw fleet — fetching URLs, parsing docs with liteparse, querying Ollama, and writing structured markdown findings. Use when doing multi-source research, deep-diving GitHub repos, Datadog docs, or any URL list. Covers the full pipeline: curl → liteparse → ollama → research/*.md

Session, layout, and pane management for Zellij terminal multiplexer. Use when managing Zellij sessions, layouts, or automating terminal workflows.

Use when a project needs role-based team coordination — PM breakdown, parallel specialist execution, review gates. Best for multi-deliverable projects vs tiger-team (single problem) or zellij-team (independent tasks).

Run Auto-Claude AI-powered ideation for any project directory. Generates low-hanging fruit, high-value features, and UI/UX ideas using the Auto-Claude backend. Use when brainstorming new specs, features, or architectural directions for a project.

Check Auto-Claude kanban card status for any project without the GUI. Shows all specs with [OK] (complete), [--] (pending), [>>] (in-progress) status. Use when checking what specs are pending, complete, or running in any Auto-Claude project.

Run Auto-Claude AI-powered roadmap generation for any project directory. Analyzes project structure, target audience, and existing specs to produce a strategic feature roadmap. Use when planning milestones, phases, or strategic direction for a project.

APFS space efficiency techniques for macOS developers. Use when copying files, cloning card bundles, creating build sandboxes, or writing Rust code that copies directories on macOS. Trigger phrases: "copy-on-write", "clonefile", "cp -c", "COW clone", "APFS", "ditto vs cp", "sparse file", "bop new card bundle". Covers clonefile() syscall, cp -c, ditto, sparse files, disk images, and snapshots.

Auto-Tundra TUI agent interface. Use when you need to inspect project state, navigate tabs, query beads/agents/sessions, or automate TUI interactions from an agent context. The TUI can run headless for programmatic access.

Autonomous Goal-directed Iteration. Apply Karpathy's autoresearch principles to ANY task. Loops autonomously — modify, verify, keep/discard, repeat. Supports optional loop count via Claude Code's /loop command.

Use when the user wants to brainstorm, ideate, or explore ideas for a topic. Creates an ideation card and optionally dispatches it to an agent. Trigger words — brainstorm, ideate, ideation, bstorm, explore ideas.

O(N) design rule for bop loops. Use when adding cross-cutting behavior (logging, hooks, events, metrics) to dispatcher, merge-gate, or any per-card loop. Trigger words — performance, loop overhead, batch, per-card cost, O(2N).

Dogfood orchestrator for working on the bop repo itself. Use when working on the bop jobcard system — create cards before coding, execute through dispatcher flow, and verify system output.

Use when explaining how to use bop from scratch — bare minimum setup, creating cards, dispatching agents, and reading results. Good for onboarding a new user who just has a project folder.

Run N opposing expert agents on a problem, each with a distinct persona and adversarial mandate. Synthesize votes, dissents, and blockers into a verdict. Trigger phrases: "get multiple opinions", "debate this", "stress-test this", "devil's advocate", "three perspectives", "adversarial review", "council". Use when reviewing specs, architecture decisions, or design contradictions. Not for code review (use pr-review-toolkit) or merge conflicts.

Backup and restore OpenCode skills to Datadog notebooks. Use when user mentions backup skills, datadog notebook, or restore skills.

Wave-based spec dispatcher for any project using dispatch.codex.nu (Codex+AC hybrid) or dispatch.nu (Auto-Claude only). Spawns Zellij panes, tracks state in JSON, pre-plans via Ollama cloud, routes specs to Codex CLI or Auto-Claude by agent field. Use when checking wave status, launching specs, rescuing stalled panes, resetting failed specs, or bootstrapping a new project dispatcher.

Generate and view a Gantt timeline of bop card runs — ANSI terminal (default) or HTML. Use when discussing execution history, parallelism, throughput, or agent performance. Trigger words — gantt, timeline, parallelism, throughput, duration, how long, when did.

Use when working in a jj (Jujutsu) repository. Covers the core workflow, workspace management, and common operations. jj is a newer VCS — do not assume git commands work here.

Use when working in the jobcard repository, dispatching agents, writing specs, creating cards, or understanding the card lifecycle. This is a multi-agent task orchestration system — NOT the transit data standard.

Leptos 0.7 CSR/WASM frontend development guide. Component patterns, signals, view! macro syntax, API calls, i18n, and 25 gotchas for building Leptos WASM apps. Use when writing or reviewing Leptos frontend code.

Deploy and manage Ollama LLM on fleet hosts. Use when user mentions ollama, LLM, AI model, or language model deployment.

Pre-dispatch quota check. Run before delegating tasks to Codex, Auto-Claude, or any agent backend. Shows live utilization for all providers and recommends which backend to use based on remaining quota. Warns when any provider is above 70% and blocks dispatch recommendations when above 90%.

Unified AI provider quota monitor. Shows Claude Code, Codex CLI, and Gemini CLI OAuth rate-limit utilization bars with reset countdowns. Reads credentials from each tool's local auth files — no API keys required.

Pre-delegation quota check. Run before dispatching Auto-Claude specs or bop cards. Shows all provider headroom, recommends the best provider, warns at >80%, and blocks dispatch suggestions at >90%. Use when about to delegate a task, start a wave, or queue specs. Trigger phrases: "can we dispatch", "quota check", "headroom", "before we delegate", "what has capacity", "which provider", "are we ok to run".

Manage Raspberry Pi fleet - OS updates, Python upgrades, ARM64-specific configurations. Use when user mentions pi, raspberry, rpi, ubuntu on pi, or ARM64.

Use when printing, logging, or verifying any credential, API key, token, or secret — pipe through redact to show only first/last 4 chars (e.g. fc1d...4439) instead of the full value

Use the local SearXNG instance when WebSearch is rate-limited, geo-blocked, or returns no results.

Use when facing a hard single problem that benefits from multiple expert perspectives simultaneously — debugging, architecture decisions, security review, incident response. Spawns 2-4 specialist agents (attacker/defender/architect/pragmatist) attacking the SAME problem from different angles.

Use when choosing Unicode symbols, icons, or glyphs for card identity tokens, terminal pane titles, filenames, UI labels, or priority/status encoding. Trigger phrases: "which symbol", "pick an icon", "glyph for", "unicode for", "what emoji", "token for", "BMP safe", "playing card suit". Covers playing cards, geometric shapes, special marks, BMP vs SMP safety, font recommendations, and filename pitfalls.

Use when working on the visual card UI, Quick Look previews, Zellij pane display, card glyphs, or the overall kanban UX of the jobcard system.

Use when spawning parallel Claude agent workers in visible Zellij panes — multi-agent visual coordination with real-time observation and ctrl-c control. Replaces tmux-based agent teams.

ZFS (OpenZFS) space efficiency and snapshotting for FreeBSD and Linux developers. Use when cloning datasets, creating build sandboxes, replicating storage, or writing Rust/C code that works with ZFS pools on macOS, FreeBSD, or Linux. Trigger phrases: "zfs snapshot", "zfs clone", "COW dataset", "zpool", "send/receive", "zfs dedup", "OpenZFS", "FreeBSD storage", "build rollback on ZFS". Covers: clone, snapshot, send/receive, dedup, Rust patterns, FreeBSD vs Linux differences.