swamp

Operates the swamp CLI to manage models, workflows, data, vaults, and extensions.

595|43|Updated Jan 28, 2026
One-click install
npx skills add https://github.com/systeminit/swamp --skill swamp
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: swamp
Source: https://github.com/systeminit/swamp/tree/main/.claude/skills/swamp
Command: npx skills add https://github.com/systeminit/swamp --skill swamp

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Working with the swamp CLI involves many subcommands across models, workflows, data, vaults, reports, extensions, and serving, and guessing flags or bypassing the CLI leads to broken definitions and stale state. This Skill routes each request to the correct operational guide so commands are run accurately and safely.

Core Features & Use Cases

  • Intent Routing: A routing table maps user intent to one of fourteen domain guides covering models, workflows, data, vaults, reports, repos, extensions, publishing, issues, sharing, serving, architecture, and troubleshooting.
  • Safe Execution Workflow: Enforces validate-before-run practices, such as running swamp workflow validate before execution and inspecting model state before destructive methods.
  • Use Case: Ask the agent to inventory EC2 instances across regions and flag untagged resources; it loads the model guide, searches for a matching model type, runs the method, and queries the resulting versioned data with CEL expressions.

Quick Start

Ask the agent to create and run a swamp workflow that checks your servers are online, and it will route to the right guide and execute the correct swamp commands.

Frequently Asked Questions about swamp

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create and run a swamp workflow?

Run swamp workflow create <name> --json to generate a scaffold with an assigned ID, edit the YAML at the returned path, then run swamp workflow validate before swamp workflow run. Never write workflow YAML from scratch or modify the id field.

How do I query model data in swamp?

Use swamp data query with a CEL predicate, such as swamp data query 'tags.env == "prod"'. Shortcuts like swamp data get, list, and versions map to equivalent queries, and all commands support --json for structured output.

When should I build a swamp model versus a workflow?

Use a model for a single external action on one resource with typed methods like create or destroy. Use a workflow when orchestrating multiple model methods in a DAG with dependencies, conditions, and parallelism.

How do I store secrets for swamp models?

Create a vault with swamp vault create <type> <name>, then store secrets with swamp vault put <vault> KEY. Model definitions reference secrets at runtime through vault expressions, and vaults can migrate between backends with swamp vault migrate.

What should I do when a swamp command fails?

First check for version drift with swamp update --check and swamp extension outdated. Then follow the four diagnostic tiers: doctor health checks, error inspection via stderr and --json output, OpenTelemetry tracing for slowness, and source fetching for internals.

When should I not use this swamp skill?

Do not use it for onboarding (use swamp-getting-started), pull requests, git operations, worktree management, cron or agent scheduling, or general coding tasks unrelated to swamp. It covers only swamp CLI operations and extension development.