Skip to main content
ClaudeWave
ES, Spanish
Slash Command67.9k estrellas del repoactualizado 5d ago

workflow-claude-subagents

This workflow slash command coordinates automated monitoring of the claude-code-best-practice project's Subagents Reference documentation by launching a research agent to detect changes in two specific areas: official frontmatter fields and built-in subagents. Use this command when you need to identify documentation drift between the official Claude Code API docs and the project's local reference document, generating a prioritized action report that tracks new, recurring, and resolved issues across multiple versions.

Instalar en Claude Code
Copiar
mkdir -p ~/.claude/commands && curl -fsSL https://raw.githubusercontent.com/shanraisshan/claude-code-best-practice/HEAD/.claude/commands/workflows/best-practice/workflow-claude-subagents.md -o ~/.claude/commands/workflow-claude-subagents.md
Después abre una sesión nueva de Claude Code; el slash command carga automáticamente.

workflow-claude-subagents.md

# Workflow Changelog — Subagents Report

You are a coordinator for the claude-code-best-practice project. Your job is to launch a research agent, wait for its results, and present a report about drift in the **Subagents Reference** report (`best-practice/claude-subagents.md`).

This workflow checks for exactly **two types of drift**:
1. **Frontmatter fields** — any field added or removed in the official docs
2. **Official sub-agents** — any built-in agent added or removed

**Versions to check:** `$ARGUMENTS` (default: 10 if empty or not a number)

This is a **read-then-report** workflow. Launch the agent, merge findings, and produce a report. Only take action if the user approves.

---

## Phase 1: Launch Research Agent

Spawn the `workflow-claude-subagents-agent` with this prompt:

> Research the claude-code-best-practice project for subagents report drift. Check the last $ARGUMENTS versions (default: 10).
>
> Fetch these 2 external sources:
> 1. Sub-agents Reference: https://code.claude.com/docs/en/sub-agents
> 2. Changelog: https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md
>
> Then read the local report (`best-practice/claude-subagents.md`).
>
> Check for exactly two things:
> 1. **Frontmatter fields**: Compare the official docs' supported frontmatter fields table against the report's Frontmatter Fields table. Flag any fields that were added or removed.
> 2. **Official sub-agents**: Compare the official docs' built-in subagents list against the report's official agents table. Flag any agents that were added or removed.

---

## Phase 2: Read Previous Changelog Entries

**While the agent is running**, read `changelog/best-practice/claude-subagents/changelog.md` to get the last 25 entries. Parse the priority actions to identify:
- **Recurring items** — issues that appeared before and are still unresolved
- **New items** — issues appearing for the first time
- **Resolved items** — previously flagged issues now fixed

---

## Phase 3: Generate Report

**Wait for the agent to complete.** Produce a report with these sections:

1. **Frontmatter Field Changes** — Fields added or removed in official docs vs our report
2. **Official Sub-agent Changes** — Built-in agents added or removed vs our table

End with a prioritized **Action Items** summary table. Each item must include a `Status` column showing `NEW`, `RECURRING (first seen: <date>)`, or `RESOLVED`:

```
Priority Actions:
#  | Type           | Action                              | Status
1  | New Field      | Add <field> to frontmatter table    | NEW
2  | Removed Field  | Remove <field> from table           | RECURRING (first seen: <date>)
3  | New Agent      | Add <agent> to official agents table | NEW
4  | Removed Agent  | Remove <agent> from table           | NEW
```

Also include a **Resolved Since Last Run** section listing items from previous runs that are no longer issues.

---

## Phase 3.5: Append Summary to Changelog

**This phase is MANDATORY — always execute it before presenting the report to the user.**

Read the existing `changelog/best-practice/claude-subagents/changelog.md` file, then **append** (do NOT overwrite) a new entry at the end. The entry format must be exactly:

```markdown
---

## [<YYYY-MM-DD HH:MM AM/PM PKT>] Claude Code v<VERSION>

| # | Priority | Type | Action | Status |
|---|----------|------|--------|--------|
| 1 | HIGH/MED/LOW | <type> | <action description> | <status> |
| ... | ... | ... | ... | ... |
```

**Status format — MUST use one of these three formats:**
- `COMPLETE (reason)` — action was taken and resolved successfully
- `INVALID (reason)` — finding was incorrect, not applicable, or intentional
- `ON HOLD (reason)` — action deferred, waiting on external dependency or user decision

The `(reason)` is mandatory and must briefly explain what was done or why.

**Rules for appending:**
- Always append — never overwrite or replace previous entries
- The date and time is when the command is executed in Pakistan Standard Time (PKT, UTC+5); get it by running `TZ=Asia/Karachi date "+%Y-%m-%d %I:%M %p PKT"`. The version comes from agent findings
- If `changelog/best-practice/claude-subagents/changelog.md` doesn't exist or is empty, create it with the Status Legend table (see top of file) then the first entry
- Each entry is separated by `---`
- **Only include items with HIGH, MEDIUM, or LOW priority** — omit NONE priority items

---

## Phase 3.6: Update Last Updated Badge

**This phase is MANDATORY — always execute it immediately after Phase 3.5, before presenting the report.**

Update the "Last Updated" badge at the top of `best-practice/claude-subagents.md`. Run `TZ=Asia/Karachi date "+%b %d, %Y %-I:%M %p PKT"` to get the time, URL-encode it (spaces to `%20`, commas to `%2C`), and replace the date portion in the badge. Also update the Claude Code version in the badge if it has changed.

**Do NOT log badge updates as action items in the changelog or report.** Badge syncing is a routine part of every run, not a finding.

---

## Phase 4: Offer to Take Action

After presenting the report (and confirming both changelog and badge were updated), ask the user:

1. **Execute all actions** — Apply all changes
2. **Execute specific actions** — User picks which numbers to execute
3. **Just save the report** — No changes

When executing:
- **New fields**: Add to the Frontmatter Fields table with correct type, required status, and description from the official docs
- **Removed fields**: Confirm with user before removing
- **New agents**: Add to the official agents table with correct #, name, model, tools, and description
- **Removed agents**: Confirm with user before removing

---

## Critical Rules

1. **Never guess** versions or dates — use data from the agent
2. **Cross-reference field counts** — report field count must match official docs
3. **Cross-reference agent counts** — report agent count must match official docs
4. **Don't auto-execute** — always present the report first
5. **ALWAYS ap
agent-browserSkill

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

agent-collectionsSlash Command

Update the AGENT COLLECTIONS table by researching all agent-collection repos in parallel

code-reviewerSubagent

Meticulous, constructive reviewer for correctness, clarity, security, and maintainability.

constitutional-validatorSubagent

Validates roadmap items, features, and technical decisions against the project's constitution, principles, and core values. Ensures all proposals align with the mission, established methodology, and design principles before implementation proceeds.

development-workflowsSlash Command

Update the DEVELOPMENT WORKFLOWS table by researching all 11 workflow repos in parallel

development-workflows-research-agentSubagent

Research agent that fetches GitHub repos, counts agents/skills/commands, gets star counts, and analyzes Claude Code workflow repositories

documentation-analyst-writerSubagent

Use this agent when you need to analyze existing documentation and create new or updated documentation that strictly adheres to project-specific documentation standards defined in claude.md. This agent excels at maintaining consistency with established documentation patterns, ensuring technical accuracy, and producing clear, well-structured documentation.

implementSlash Command

Execute phased implementation with validation gates