Skip to content
Back to skills

Workflows Workflow Execute

ASecurity

Run multi-agent workflows — in the conversation with Task-tool agents, or as a persistent org with npx monomind org run

  • 21 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 10, 2026
ai-agentsjavascriptgojavabashapi

Works with

  • claude code
  • cli
  • api

Security analysis

A100/100

Scanned October 1, 2026

npx -y skills add monoes/monomind --skill workflows-workflow-execute --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Workflows Workflow Execute?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Workflows Workflow Execute
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/monoes-workflows-workflow-execute/badge)](https://www.skillsdirectory.com/skills/monoes-workflows-workflow-execute)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: workflows-workflow-execute
description: Run multi-agent workflows — in the conversation with Task-tool agents, or as a persistent org with npx monomind org run
type: flow
---

# Workflow Execute

Run a multi-stage, multi-agent workflow. The CLI has no `workflow` command; use one of the two real paths below.

## How to Invoke

```
Skill("workflows:workflow-execute")
```

---

## Path 1: In the Conversation (Task Tool)

Claude Code executes the stages directly:

1. Break the work into stages with TodoWrite (plan → implement → test → review)
2. Spawn every independent agent for a stage in ONE message via the Task tool
3. Wait for results, review them, then start the next stage

```javascript
// Stage: implement + test in parallel
Task({ subagent_type: "coder",  prompt: "Implement the auth API in src/auth/ per the plan ..." })
Task({ subagent_type: "tester", prompt: "Write tests for the auth API contract ..." })

// Next stage: review
Task({ subagent_type: "reviewer", prompt: "Review the diff in src/auth/ ..." })
```

Optionally track the task for routing:

```bash
npx monomind hooks pre-task --description "Build REST API with auth" --task-id api-auth
# ... stages run ...
npx monomind hooks post-task --task-id api-auth --success true
```

## Path 2: As an Org (`monomind org`)

For repeatable, scheduled, or long-running workflows:

```bash
# Create from a starter template (or use an existing .monomind/orgs/<name>.json)
npx monomind org create api-dev --template dev-team --goal "Build REST API with auth"

# Validate and preview each role's briefing without starting agents
npx monomind org validate api-dev
npx monomind org run api-dev --dry-run

# Run (foreground daemon); override the goal for this run
npx monomind org run api-dev --task "Add refresh-token rotation"

# Cap spend
npx monomind org run api-dev --budget-usd 5
```

### `org run` Flags

| Flag | Description |
|------|-------------|
| `--task` | Override the org goal for this run |
| `--dry-run` | Validate and print each role's briefing without starting sessions |
| `--resume` | Resume from the persisted checkpoint instead of starting fresh |
| `--budget-usd` | Hard-stop if the upfront cost estimate exceeds this value |
| `-y, --yes` | Skip the cost-estimate confirmation prompt |

## After Launch

```bash
npx monomind org status api-dev          # runtime state
npx monomind org logs api-dev --follow   # live event log
npx monomind org questions api-dev       # pending ask_human questions
npx monomind org approvals api-dev       # pending tool approvals
npx monomind org pause api-dev           # finish current turns, start no new cycles
npx monomind org stop api-dev            # stop the daemon
npx monomind org report api-dev          # outcome, per-role activity, tokens
```

## Related Skills

- `workflows:workflow-create` — Create a reusable workflow as an org config
- `workflows:development` — Development workflow pattern
- `workflows:research` — Research workflow pattern
- `mastermind-runorg` — Start a saved org

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.