Irrlicht

End-to-end execution of a GitHub issue via `/ir:exec [mode] <N>` — open a worktree, investigate, then either present a visual HTML plan for approval or proceed straight to implementation (mode-dependent), open a PR, review it with /code-review, fix findings, simplify, hand back the PR link with a test-or-merge recommendation, and land the merge on request. `mode` defaults to `auto`, which picks `plan` or `full` from the issue's `/ir:triage` signals. Triggers on "/ir:exec", "exec issue", "implement issue", "plan issue", or when the user gives an issue number/URL and asks to plan, implement, or land it.

Build and publish an irrlicht release. Bumps version, builds Go daemon + Swift app, creates signed app bundle with icon, packages DMG (branded installer) + PKG, updates docs/changelog/landing page, commits, tags, pushes, and creates GitHub release with assets. Default: patch bump. Use "/ir:release minor" or "/ir:release major".

Audit all project documentation for understandability, completeness, and validity using objective, binary criteria, then fix every finding directly in place. Filing a GitHub issue is the rare fallback — only for a finding whose fix is genuinely ambiguous or that fails verification after being applied. Every finding is anchored to a quoted location and stamped with a stable ID so independent runs converge. Triggers on "/ir:doc-review", "audit docs", "review documentation", "check the docs", "doc audit". Supports a --report-only dry run that changes nothing and files nothing.

Check latest releases of coding agents monitored by irrlicht (Claude Code, OpenAI Codex, Pi, Gas Town, Aider, OpenCode, Gemini CLI, Kiro CLI, Antigravity, Mistral Vibe) and report new features that impact session monitoring. Use when user says 'agent releases', 'check releases', 'agent updates', '/ir:agent-releases', or wants to know if upstream agent changes affect irrlicht.

Autonomously triage GitHub issues on ingo-eichhorst/Irrlicht. Diagnostic-only: scores each issue against a 7-axis readiness rubric and lands it at ready-for-agent or needs-info, with a one-line justification per label. Triggers: "/ir:triage" (sweep), "/ir:triage #N" (single), "/ir:triage #N #M".

Build and run a dev irrlicht daemon and/or macOS Swift app for local testing. Two independent axes: MODE (replace, default — takes over production on port 7837/production state; separate — coexists on port 7838/isolated state) and TARGET (full, default — daemon + app; daemon — just rebuild+restart irrlichd, leave the running app to reconnect; macos — just rebuild+relaunch the Swift app against whatever daemon is already up). Use when the user says "test mac", "restart mac", "rebuild mac", "restart just the daemon", "rebuild the mac app only", or "/ir:test-mac".

Judge one (agent, scenario) cell across the three pillars — agent capability, daemon sensor capture, driver capability — on cited evidence, then author the cell's recipe and machine-checkable spec. Writes the cell metadata via `of cell write` and the spec (expected.jsonl) via `of cell spec`. No live recording. Invoked as `/ir:onboarding-factory assess <agent> <scenario>`.

Carry one assessed cell to a committed, verified recording: check prerequisites, port any missing driver step, drive the live agent CLI under a recording daemon via `of record run`, verify EVERY websocket observation (state + model + cost + tokens + agent) via `of record verify`, refresh the replay golden, and commit. Backflows a correction into the cell when the live recording disagrees with the assessment. Invoked as `/ir:onboarding-factory record <agent> <scenario>`.

Maintain the canonical scenario × agent fixture matrix for irrlicht. A slim dispatcher that routes intent to four focused subagents — `create-scenario` (add a matrix row), `create-agent` (add a matrix column), `assess` (judge one cell across the three pillars and write its spec), and `record` (drive the live agent and verify every websocket observation). Every read and every write goes through the `of` factory CLI (tools/onboarding-factory) — the skill itself never touches `replaydata/`. Each subagent returns a short summary (≤6 lines, ≤7 for `record`) so the parent keeps its context for strategic decisions instead of drowning in per-cell tool output. Use when the user says "/ir:onboarding-factory", "onboard agent", "add a scenario", "assess fixtures", "record fixtures", or "regenerate recordings". For unattended / overnight runs ("push through", "onboard everything", "implement everything that can be"), see the Overnight push-through mode section — implement every feasible unlock and deliver a report-as-PR,

Scan the web for coding agents and agent orchestrators, track GitHub stars and trends, rank by popularity+momentum, and publish a report to the irrlicht site. Shows which agents irrlicht already supports. Use when user says 'agent landscape', 'scan agents', 'coding agent tracker', 'agent popularity', '/ir:agent-landscape', or wants to see the competitive landscape of coding agents.

Sync irrlicht's model-name alias map against codeburn's upstream BUILTIN_ALIASES. Fetches codeburn's src/models.ts, diffs entries against core/pkg/capacity/aliases.go, and proposes additions/changes as a PR. Use when user says '/ir:refresh-aliases', 'refresh aliases', 'sync codeburn aliases', 'check alias map', or when a session prices at $0 for a known model.

On-demand SonarQube Cloud (SonarCloud) code-quality report — fetches open issues for this repo's connected project, sorted worst-first, with concrete file:line locations, rule keys, fix guidance, source-to-sink taint traces for security findings, and a Security Hotspots check. Runs entirely locally via tools/sonarqube-report.sh (SONAR_TOKEN lives in a local .env, not a CI secret). Use when the user says '/ir:sonarqube-report', 'run a sonarqube report', 'sonar report', 'what needs fixing', 'show me the sonarqube issues', or wants concrete, actionable code-quality findings (as opposed to /ir:codescene-report's hotspot/trend view, which doesn't exist anymore).

Use this skill to generate well-branded interfaces and assets for Irrlicht, either for production or throwaway prototypes/mocks/etc. Contains essential design guidelines, colors, type, fonts, assets, and UI kit components for prototyping.

Onboard a brand-new agent CLI as a matrix COLUMN: research its identity + recording prerequisites, register it via `of agent add`, scaffold its interactive driver from the template, and predict which step types each scenario's recipe will need (the driver-needs punch-list). No live recording. Invoked as `/ir:onboarding-factory create-agent <slug>`.

Add a brand-new scenario ROW to the matrix: one agent-agnostic `{id, name, description, acceptance_criteria, process}` entry written through `of scenario add`. Researches how the behavior manifests across every onboarded agent (and what the daemon would observe) before synthesizing the agent-agnostic spec. No agent CLI invocation, no recording. Invoked as `/ir:onboarding-factory create-scenario <slug>`.