Menu
Research and catalog workflows from external Claude Code repositories
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Migrate a single-harness repo to the dual-harness layout so both Claude Code and Codex load the same instructions and skills — AGENTS.md as the portable canonical doc, a thin CLAUDE.md `@AGENTS.md` bridge, and a `.agents/skills` dir with a `.claude/skills` symlink onto it. Use when asked to migrate to dual-harness, make this repo work in both Claude and Codex, or for agent metadata cleanup.
Diagnose and fix Catalyst setup issues. Validates tools, database, config, OTel, direnv, and thoughts. Automatically fixes what it can — creates directories, initializes the database, sets WAL mode, runs migrations. Use for new installs, upgrades, or when something isn't working.
Fast browser automation CLI for AI agents. **ALWAYS use instead of Playwright MCP tools** for web testing, screenshots, form filling, and UI verification. Use when user says "open in browser", "check the site", "take a screenshot", "fill the form", "test the UI", or any browser interaction. Also use as a fallback when a task requires visual browser interaction that CLIs and APIs cannot handle (e.g., OAuth flows, complex dashboards, visual verification).
Goal-driven senior-engineer pipeline-unstick sweep (CTL-1176 rung 3). Given the stuck/failed/needs-human set (or ONE ticket handed by the recovery router), its GOAL is to get the pipeline MOVING again — not to fix one ticket's review findings (that is phase-remediate). It runs AFTER the eyes (diagnostician evidence) and the hands (deterministic unstuck-sweep seams) have already tried, and it CONSUMES their output from a recovery-pass.json brief rather than re-diagnosing or redoing their narrow work. It acts like a senior engineer with full tool access — it resolves merge conflicts, rebases, force-pushes, merges green PRs, and re-dispatches stalled phases AUTONOMOUSLY — and escalates to the operator ONLY for a genuine value judgment / something that degrades other functionality / a real cost-benefit trade-off / a serious architecture change / an ADR conflict. On escalation it AUTHORS the operator inbox row + the push notification (executive-voiced). Dispatched as a `claude --bg` job by phase-agent-dispatch via
Systematically pull, categorize, and address all PR review comments — code change requests, questions, and suggestions. This skill fetches comments via gh api, groups them by file, implements fixes, handles disagreements diplomatically, and pushes a single commit. You should not try to handle PR review feedback manually — this skill ensures nothing gets missed. **ALWAYS consult this skill when** the user says 'address comments', 'fix review feedback', 'handle PR comments', 'respond to reviewers', 'address review', 'review feedback', or mentions that a PR has unresolved comments or review threads. Also used by /oneshot Phase 5 to process reviewer feedback before merging.
Reference template every phase-agent skill copies (CTL-448). The leading underscore on the directory name prevents the skill loader from picking this up — it is NOT a runnable skill, only a structural template the nine real phase agents (phase-triage, phase-research, phase-plan, phase-implement, phase-verify, phase-review, phase-pr, phase-monitor-merge, phase-monitor-deploy) clone and specialize. The real phase skills MUST set `user-invocable: true` so `phase-agent-dispatch`'s `claude --bg "/catalyst-dev:phase-X ..."` slash command resolves (CTL-490).
| name | discover-workflows |
| description | Research and catalog workflows from external Claude Code repositories |
| disable-model-invocation | false |
| allowed-tools | Read, Write, Task, TodoWrite, WebSearch, WebFetch |
| version | 1.0.0 |
You are tasked with researching external Claude Code repositories to discover, analyze, and catalog their agents, commands, and workflow patterns.
This command helps you learn from the Claude Code community by analyzing workflow repositories and extracting reusable patterns.
Default repositories to research:
catlog22/Claude-Code-Workflow - Multi-agent automationautomazeio/ccpm - Project management systemwshobson/commands - Production slash commandswshobson/agents - Production subagentsqdhenry/Claude-Command-Suite - 148+ commands, 54 agentsVoltAgent/awesome-claude-code-subagents - 100+ subagentshesreallyhim/awesome-claude-code - Curated commands/agentsfeiskyer/claude-code-settings - Workflow improvementsOneRedOak/claude-code-workflows - Code review workflowsanthropics/claude-code - Official Claude Code repowinfunc/opcode - GUI toolkit for agents/commandsWhen invoked:
I'll research Claude Code workflows from external repositories.
Which repository would you like to explore?
1. wshobson/commands - Production slash commands
2. wshobson/agents - Production subagents
3. qdhenry/Claude-Command-Suite - 148+ commands
4. VoltAgent/awesome-claude-code-subagents - 100+ subagents
5. Custom repository (provide org/repo)
Or type 'all' to catalog all supported repos (this may take a while).
Get user selection or use provided parameter.
IMPORTANT: Spawn 3 parallel research tasks for efficiency and context isolation.
Use TodoWrite to track the 3 parallel research tasks.
Task 1 - Workflow Discovery:
Use external-research agent:
"Research {repo-name}. What commands and agents are available? List all workflows with brief descriptions of what each does."
Tools: WebFetch, WebSearch
Return: Complete list of all workflows found
Task 2 - Frontmatter Analysis:
Use external-research agent:
"Research {repo-name}. What frontmatter format is used for agents and commands? Provide specific examples showing all frontmatter fields used."
Tools: WebFetch, WebSearch
Return: Frontmatter patterns with concrete examples
Task 3 - Implementation Patterns:
Use external-research agent:
"Research {repo-name}. What are the common implementation patterns, structures, and conventions used across workflows? Include naming conventions, file organization, and any templates."
Tools: WebFetch, WebSearch
Return: Patterns, templates, conventions observed
WAIT for all 3 tasks to complete before proceeding.
Why parallel:
Combine findings from the 3 parallel research tasks:
Mark all 3 tasks complete in TodoWrite.
From the aggregated results, extract:
Available Workflows
Frontmatter Patterns
Implementation Patterns
Unique Features
Save research to thoughts/shared/workflows/{repo-name}/analysis.md:
# Workflow Analysis: {Repo Name}
**Repository**: {org/repo} **Analyzed**: {date} **Focus**: {agents/commands/both}
## Summary
[1-2 sentence overview of what this repo offers]
## Available Workflows
### Commands
1. **{command-name}**
- **Purpose**: [what it does]
- **Use when**: [scenario]
- **Frontmatter**:
```yaml
[actual frontmatter from repo]
```
2. **{command-name}** [...]
### Agents
1. **{agent-name}**
- **Purpose**: [what it does]
- **Tools**: [tools it uses]
- **Frontmatter**:
```yaml
[actual frontmatter from repo]
```
## Frontmatter Patterns
### Standard Fields
- name: [how they define it]
- description: [format they use]
- tools: [how specified]
- [other fields observed]
### Naming Conventions
- [pattern 1]
- [pattern 2]
## Implementation Patterns
### Common Structures
[Patterns you notice across workflows]
### Reusable Templates
[Templates that could be adapted]
## Unique Features
[Novel or interesting approaches]
## Integration Notes
[How these could integrate with your workspace]
## Recommendations
### High-Value Imports
1. **{workflow-name}** - [why it's valuable]
2. **{workflow-name}** - [why it's valuable]
### Patterns to Adopt
- [Pattern 1]: [how to use it]
- [Pattern 2]: [how to use it]
## References
- Repository: {URL}
- Analyzed on: {date}
Update thoughts/shared/workflows/catalog.md:
# Workflow Catalog
Discovered workflows from the Claude Code community.
## Repositories Analyzed
### wshobson/commands
- **Analyzed**: 2025-01-08
- **Workflows**: 15 commands
- **Focus**: Production-ready automation
- **Details**: [See analysis](wshobson-commands/analysis.md)
- **Top Picks**:
- code-review: Automated code review workflow
- refactor: Safe refactoring patterns
[... more repos]
## By Category
### Code Review
- wshobson/commands: code-review
- OneRedOak/claude-code-workflows: review-pr
### Documentation
- qdhenry/Claude-Command-Suite: doc-generator
- hesreallyhim/awesome-claude-code: readme-generator
[... more categories]
## By Use Case
### "I want to automate code reviews"
1. wshobson/commands/code-review
2. OneRedOak/claude-code-workflows/review-pr
3. [Details in respective analyses]
### "I need project management workflows"
1. automazeio/ccpm - Full PM system
2. [...]
Show user what was found:
# Discovery Results: {Repo Name}
## Summary
Discovered {N} workflows ({X} commands, {Y} agents)
## Highlights
### Top Workflows
1. **{name}** - {brief description}
2. **{name}** - {brief description}
3. **{name}** - {brief description}
### Interesting Patterns
- {Pattern 1}
- {Pattern 2}
### Recommended for Import
- **{workflow-name}**: {why}
## Next Steps
1. **Review the analysis**: `thoughts/shared/workflows/{repo}/analysis.md`
2. **Import a workflow**: `/catalyst-meta:import-workflow {repo} {workflow-name}`
3. **Discover another repo**: `/catalyst-meta:discover-workflows`
Catalog updated at: `thoughts/shared/workflows/catalog.md`
/catalyst-meta:discover-workflows all
This will:
Performance: ~10-15x faster than sequential research
Context efficiency: Each agent loads only its research area
/catalyst-meta:discover-workflows org/repo
Works with any public GitHub repo with Claude Code workflows.
/catalyst-meta:discover-workflows wshobson/agents --focus agents
Only analyzes agents, skips commands.
/catalyst-meta:import-workflow to actually import/catalyst-meta:discover-workflows (this command)/catalyst-meta:import-workflow (imports discovered workflows)/catalyst-meta:create-workflow (creates new using discovered patterns)/catalyst-meta:validate-frontmatter (ensures consistency)This command is the first step in workflow discovery and reuse!