Stats
Actions
Tags
From pace
Use when the user invokes a /pace sub-command, or wants a company-standard way to do work across engineering, product, ops, marketing, or other functions. Provides shared vocabulary and repeatable processes between humans and AI coding agents. Not for one-off code tasks that don't match a sub-command.
How this skill is triggered — by the user, by Claude, or both
Slash command
/pace:pace [command] [target][command] [target]This skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Pace is the GoldenBerry company-wide skill router. One entry point (`/pace`), many sub-commands, each one a repeatable way of working that humans and AI agents can share.
Pace is the GoldenBerry company-wide skill router. One entry point (/pace), many sub-commands, each one a repeatable way of working that humans and AI agents can share.
For frontend design specifically, pace defers to impeccable: install it alongside pace and call /impeccable <command> for design work. Pace covers everything else: engineering review, product specs, ops runbooks, marketing copy, and more as commands land.
Before doing any work in a sub-command:
reference/<command>.md. Skipping this produces generic output that ignores the company's way of working.node .claude/skills/pace/scripts/load-context.mjs
Consume the full JSON output. The output's contextDir field tells you where the file was resolved from. If PRODUCT.md is missing, proceed but flag that pace works better with a project brief.
Sub-commands live in reference/<command>.md. The router below is the source of truth; bun run build reads it to wire up the harness output dirs.
| Command | When to use |
|---|---|
cleanup | Engineering housekeeping: safely sweep stale dev docs/artifacts, merged branches, and closeable issues/tickets left over from old development. |
When the user types /pace <command> and the command isn't in the table above, respond: "Pace doesn't have a <command> command yet. Want to author one? Drop a reference file at skill/reference/<command>.md and add a row here."
Users can promote any pace command to a standalone top-level slash command with node .claude/skills/pace/scripts/pin.mjs <command>. This creates a thin redirect so /audit (for example) delegates to /pace audit. pin.mjs --unpin <command> removes the shim.
The pin allowlist lives in pin.mjs (VALID_COMMANDS). Keep it in sync with the table above.
npx claudepluginhub goldenberry-so/pace --plugin paceProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.