hcaiano

art-director

Manual-only long design-exploration loop. Use only when the user explicitly invokes art-director: direct Codex image batches, curate and wipe between axes, and converge on a winning direction before handing it to impeccable or image-to-code.

check-logs

Check logs in an existing turborepo dev TUI inside herdr. Use when the user asks about an app's dev output, whether an app is running, or an error shown by a `bun dev` / `turbo run dev` pane. Read-only: navigate the existing TUI; never start, stop, or restart the server.

goal-loop

Manual-only long improvement campaign run by Claude + Codex as a pair. Use only when the user explicitly invokes goal-loop and supplies one target plus what 'better' means; the agents repeatedly test, improve, ship PRs, and require current evidence plus peer acceptance to stop.

herdr-orchestrate

Manual-only orchestrator for delegating GitHub issues to agent pairs in dedicated Herdr tabs: triage, work units, worktree + pair per unit, implement → ship-it, status on demand.

ship-it

Manual-only PR creation for current work: run the local dual-review gate, commit and push intentional changes, open the PR, and wait for green CI. Use only when the user explicitly invokes ship-it.

ask-peer

Opposite-model request for one focused pass. In Codex, reach for this autonomously to escalate to Fable (through `claude -p`) when you need stronger reasoning, a second opinion, or a scoped review; in Claude Code it stays manual — Claude asks Codex through `codex exec` only when the user invokes it. Use for a focused pass without a live Herdr pair.

Code & Development Solid

review-pr-comments

On-demand review cleanup for an existing PR that has actual feedback: verify each comment, apply valid in-scope fixes in one batched round, respond where useful, and leave the PR merge-ready. Only runs when the user explicitly invokes review-pr-comments.

cmux-pair

Legacy cmux-based pair programming bootstrap for Claude and Codex talking through cmux surfaces with an XML protocol. Keep this for historical reference and migration only. Prefer herdr-pair for current Claude/Codex collaboration; use this skill only when the user explicitly asks for the old cmux workflow.

cyber-audit

Read-only exposure audit of THIS MACHINE against a specific external advisory — a named CVE, malicious package, or supply-chain advisory — then write a report. Invoke only when the user explicitly asks whether THIS MACHINE is affected by a named CVE/package/advisory ("am I affected by X", "is this package/CVE on my system", "scan my machine for Y"). Do not invoke merely because an advisory/CVE/package is mentioned, for an explanation or news summary, for code/repo/appsec review, or for account/credential-breach response.

Code & Development Solid

debug-mode

Evidence debugging for unresolved bugs. Use when the user says 'diagnose' or 'debug this', or when a bug is flaky, environment-specific, production-only, or its root cause survived a normal reproduction/inspection. Build a feedback loop, collect runtime evidence, and fix only after evidence identifies the cause. Do not use for a compiler/type/lint/test error whose cause is in the output, a review finding that already names cause and fix, an already-proven root cause, or from inside an active debug-mode run.

AI & Automation Listed

herdr-pair

Persistent Claude and Codex coworking inside one Herdr tab. Use when the user asks for Herdr pairing, another workflow needs a live peer, or input begins with `[agent` or includes `[herdr-pair control`. For inbound traffic, including after context compaction, run receive and reply through the bundled sender instead of local output.