Claude Skills by smithersai

Generate AI SDK tools from OpenAPI specs. Run `smithers openapi --help` for usage details.

Run GEPA prompt optimization over a workflow eval suite and write an optimized prompt artifact. Run `smithers optimize --help` for usage details.

Print node output row. Run `smithers output --help` for usage details.

Gracefully pause a run: stop scheduling new tasks, let in-flight tasks finish, then park it resumably. Run `smithers pause --help` for usage details.

List active, paused, and recently completed runs. Run `smithers ps --help` for usage details.

Fork from a checkpoint and resume execution (time travel). Run `smithers replay --help` for usage details.

Restore a worktree to a durability checkpoint (latest for the node, or --seq). Run `smithers restore --help` for usage details.

Retry a specific task within a run, then resume the workflow. Run `smithers retry-task --help` for usage details.

Revert the workspace to a previous task attempt's filesystem state. Run `smithers revert --help` for usage details.

Run code review plus story-form HTML walkthrough generation for a repo or PR. Run `smithers review --help` for usage details.

Rewind a run to a previous frame. Run `smithers rewind --help` for usage details.

View scorer results for a specific run. Run `smithers scores --help` for usage details.

Deliver a durable signal to a run waiting on <Signal> or <WaitForEvent>. Run `smithers signal --help` for usage details.

Internal: PostToolUse hook that requests a Tier 1 durability snapshot. Run `smithers snapshot-hook --help` for usage details.

List durability snapshots (workspace checkpoints) for a run. Run `smithers snapshots --help` for usage details.

Show plain-English starter workflows with copy-paste commands. Run `smithers starters --help` for usage details.

Concise run health at a glance: verdict, node counts, agent/model mix, throughput, and the nodes gating progress. Run `smithers status --help` for usage details.

Watch explicitly named stale runs and auto-resume them; pass --all to opt into a workspace-wide sweep. Run `smithers supervise --help` for usage details.

View execution timeline for a run and its forks (time travel). Run `smithers timeline --help` for usage details.

Time-travel to a previous task state by reverting filesystem state, resetting DB state, and optionally resuming. Run `smithers timetravel --help` for usage details.

Issue and revoke short-lived Gateway bearer tokens. Run `smithers token --help` for usage details.

Print DevTools snapshot as XML tree. Run `smithers tree --help` for usage details.

Open the custom UI for a workflow run in your browser. Starts a local Gateway automatically if none is running (serving workflow-owned <UI> declarations); pass --no-autostart or --gateway <url> to opt out. Run `smithers ui --help` for usage details.

Start a workflow execution. Omit the workflow (or pass --interactive) to pick one interactively and monitor it in the full-screen TUI; use -d for detached (background) mode. Run `smithers up --help` for usage details.

Check for a newer Smithers release and upgrade the install (or print how). Run `smithers update --help` for usage details.

Run the agent-assisted Smithers upgrade workflow: fetch changelogs, upgrade with a cheap agent, and escalate to a smart agent only when needed. Run `smithers upgrade --help` for usage details.

Show how much rate limit / subscription quota each registered account has used. Run `smithers usage --help` for usage details.

Summarize what happened in a run or node: a cheap fast agent narrates the recorded facts (deterministic recap when no agent is available). Run `smithers what --help` for usage details.

Explain why a run is currently blocked or paused. Run `smithers why --help` for usage details.

Discover local workflows from .smithers/workflows. Run `smithers workflow --help` for usage details.

Inspect and reclaim the worktrees Smithers created for <Worktree> lanes. Run `smithers worktree --help` for usage details.

Drive Smithers, a durable control plane for long-running coding agents, from Claude Code. Use when the user wants multi-step, long-running, crash-safe, or human-in-the-loop agent work ('orchestrate agents', 'run a workflow', 'implement this and review it', 'keep iterating until tests pass', 'plan then build') or anything needing retries, approvals, replay, or evals across multiple AI steps. YOU (Claude) run Smithers on the user's behalf; it is not a GUI the human clicks. HARD RULE 0, if `SMIT...

Drive Smithers, a durable control plane for long-running coding agents, from Codex. Use when the user wants multi-step, long-running, crash-safe, or human-in-the-loop agent work ("orchestrate agents", "run a workflow", "implement this and review it", "keep iterating until tests pass", "plan then build") or anything needing retries, approvals, replay, or evals across multiple AI steps. YOU (Codex) run Smithers on the user's behalf; it is not a GUI the human clicks. HARD RULE 0: if `SMITHERS_IN...

The concierge proxy — turn a vague user script ("I need the agent to help me do X") into a context contract, route it to the right skills/workflows, add backpressure (tests/evals/reviews/approvals), execute, and report. Use when a request is multi-step, durable, or human-in-the-loop and you'd otherwise hand-roll the orchestration; skip it for a single prompt → single answer.

How this repo does eval-driven development (EDD) for Smithers workflows — write the failing suite first, build until green, validate on a holdout, then optimize. Use when adding evals to a workflow, changing a prompt/model/graph that has a suite, setting up a dev/holdout split, or wiring eval gates into CI. For suite-authoring mechanics see skills/eval-writer/SKILL.md; this skill is the loop and the discipline.

Turn acceptance criteria into a runnable Smithers eval suite (JSONL cases + rubric) and wire it to `smithers eval`. Use when a workflow's quality must be measured and regression-tested — not "looks good" once, but a repeatable check that fails when the model OR the harness regresses.

Design a single high-quality prompt — the innermost layer an agent reads. Use when a Smithers <Task>'s prompt (its .mdx body or inline string) is vague, ambiguous, or underperforming and you want to tighten the instruction, role, constraints, examples, and output contract before reaching for harness or workflow changes.

Turn a run's persisted state into a self-contained HTML slideshow report — objective, decisions, graph, gates, evals, artifacts, failures, and the recommended next run. Use when a long or important run needs a legible, shareable summary instead of scrolling raw logs.

Safety/risk review for agent actions — the backpressure layer that gates side effects (secrets, external publishing, destructive repo changes, expensive ops) behind human approval inside a Smithers workflow. Use when a workflow performs outward-facing or irreversible actions.

Design the Zod output schema of a Smithers <Task> as the contract between steps. Use when a step's output feeds a later step (or a branch/loop condition) and must be reliable — design the schema first, keep it minimal, and prefer typed fields over prose so downstream rendering can depend on it.

Use every registered Claude Code and Codex subscription as one quota-aware failover pool in Smithers workflows, so rate limits on a single account never stall a run. Use when the user mentions rate limits, quota, multiple subscriptions/accounts, "fallback agents", adding a new Claude/Codex account to Smithers, or wants tasks spread across their team's subscriptions. Covers: registering a new account with a tmux browser login, verifying it, and wiring `fallbackAgents()` into a workflow.

Drive Smithers, a durable control plane for long-running coding agents. Use when the user wants multi-step, long-running, crash-safe, or human-in-the-loop agent work: "orchestrate agents", "run a workflow", "implement this and review it", "keep iterating until tests pass", "plan then build", or anything that needs retries, approvals, replay, or evals across multiple AI steps. YOU (the agent) run Smithers on the user's behalf; it is not a GUI the human clicks. HARD RULE 0: if `SMITHERS_INSIDE_...

Prepare, launch, and evaluate Kimi LoRA fine-tuning for Smithers TypeScript/JSX workflow authoring. Use when building a Smithers training corpus, validating Fireworks JSONL, comparing base and tuned Kimi models, starting a Fireworks Kimi K2.7 managed SFT job, requesting or using Kimi K3 Serverless Training access, estimating training spend, or refreshing the checked-in Smithers authoring dataset.