# Send swamp to your agent Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually. ## Fast path - Download the package from Yavira. - Extract it into a folder your agent can access. - Paste one of the prompts below and point your agent at the extracted folder. ## Suggested prompts ### New install ```text I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete. ``` ### Upgrade existing ```text I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run. ``` ## Machine-readable fields ```json { "schemaVersion": "1.0", "item": { "slug": "swamp", "name": "swamp", "source": "tencent", "type": "skill", "category": "开发工具", "sourceUrl": "https://clawhub.ai/umag/swamp", "canonicalUrl": "https://clawhub.ai/umag/swamp", "targetPlatform": "OpenClaw" }, "install": { "downloadUrl": "/downloads/swamp", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=swamp", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "packageFormat": "ZIP package", "primaryDoc": "SKILL.md", "includedAssets": [ "SKILL.md", ".claude/settings.local.json" ], "downloadMode": "redirect", "sourceHealth": { "source": "tencent", "slug": "swamp", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-05-12T18:03:23.184Z", "expiresAt": "2026-05-19T18:03:23.184Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=swamp", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=swamp", "contentDisposition": "attachment; filename=\"swamp-0.2.0.zip\"", "redirectLocation": null, "bodySnippet": null, "slug": "swamp" }, "scope": "item", "summary": "Item download looks usable.", "detail": "Yavira can redirect you to the upstream package for this item.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/swamp" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] } }, "links": { "detailUrl": "https://openagent3.xyz/skills/swamp", "downloadUrl": "https://openagent3.xyz/downloads/swamp", "agentUrl": "https://openagent3.xyz/skills/swamp/agent", "manifestUrl": "https://openagent3.xyz/skills/swamp/agent.json", "briefUrl": "https://openagent3.xyz/skills/swamp/agent.md" } } ``` ## Documentation ### Swamp — API Modeling & *Claw Enrichment You are an AI agent that uses Swamp to model any API, test it, and turn it into a reusable *Claw capability. Swamp is an AI-native automation CLI that represents external resources (APIs, CLIs, cloud services) as typed models with executable methods and composable workflows. ### Repository Setup Before any operation, check if the current directory is a swamp repository. If not, initialize one: swamp repo init --tool claude This creates a .swamp/ directory with the necessary structure. Use --tool claude to set up the AI agent integration for Claude/OpenClaw. ### Step 1: Discover Model Types Find available model types to understand what kinds of resources can be modeled: swamp model type search swamp model type describe The command/shell type is the most versatile — it can model any API via shell commands (curl, httpie, CLI tools). ### Step 2: Create a Model Create a new model definition for the API: swamp model create Example: swamp model create command/shell github-issues ### Step 3: Edit the Model Definition Open and edit the model's YAML definition to configure endpoints, authentication, parameters, and methods: swamp model edit When editing, define: Methods: The operations this API supports (list, create, update, delete) Inputs: Parameters each method accepts Authentication: Reference vault secrets via CEL expressions Commands: The actual shell commands (curl calls, CLI invocations) that execute each method ### Step 4: Validate Ensure the model definition is well-formed: swamp model validate Fix any validation errors before proceeding. ### Run a Method Execute a method on your model to test it: swamp model method run ### Inspect Results Check outputs, logs, and data from the execution: swamp model output get swamp model output logs swamp model output data ### Review History See past executions: swamp model method history search swamp model method history get Iterate on the model definition until methods return the expected results. ### Workflow Orchestration Chain multiple model methods into automated workflows with dependency ordering: swamp workflow create swamp workflow edit swamp workflow validate swamp workflow run Workflows support: Parallel job execution Dependency ordering between jobs Trigger conditions via CEL expressions Cross-model references (one model's output feeds into another) Check workflow results: swamp workflow history get swamp workflow history logs ### Vault & Credentials Store API keys and secrets securely — never hardcode them: swamp vault create swamp vault put swamp vault list-keys Reference secrets in model definitions using CEL expressions like vault.get("my-vault", "API_KEY"). ### Data Management Inspect outputs and artifacts across models: swamp data list swamp data get swamp data search ### Swamp-Club Authentication Swamp-club is the community registry and collaboration layer. Authenticate to push/pull extensions and share work: swamp auth login swamp auth login --server swamp auth whoami swamp auth logout swamp auth login opens a browser-based login flow by default Use --no-browser with --username and --password for headless/CI environments Set SWAMP_CLUB_URL env var to target a custom server Always verify identity with swamp auth whoami after login Authentication is required before pushing extensions to the registry. ### Extensions Extensions expand Swamp's model types and capabilities. Use the extension registry to share and reuse custom models: ### Pull an Extension Install a community or team extension from the registry: swamp extension pull swamp extension pull --force Use --force to overwrite existing files without prompting. ### List Installed Extensions See what extensions are currently installed: swamp extension list ### Push an Extension Publish your own extension to the swamp registry (requires swamp auth login first): swamp extension push swamp extension push --dry-run Use --dry-run to build the archive locally and verify it without actually publishing. Use -y to skip confirmation prompts. ### Remove an Extension Uninstall a pulled extension and clean up its files: swamp extension remove ### Extension Workflow The typical flow for extending Swamp with new model types: Create custom TypeScript model definitions in extensions/models/ Test locally with swamp model create Package with a manifest file Dry-run: swamp extension push ./manifest.yaml --dry-run Publish: swamp extension push ./manifest.yaml Others install: swamp extension pull ### Enriching *Claw with New Capabilities Once a Swamp model is validated and working, turn it into a standalone *Claw skill: Export the model: Use swamp model get --json to extract the full definition Generate a SKILL.md: Create a new skill file that wraps the Swamp model's methods as agent instructions Include setup instructions: Document the required env vars, binaries, and vault configuration Publish to ClawHub: Share the skill with the community via clawhub publish The generated skill should: Require swamp in its bins dependency Reference the swamp repo and model by name Map each model method to a clear agent instruction Include examples of typical invocations ### Sharing via Extensions For reusable model types (not just individual models), publish as a Swamp extension: Build the custom model type in extensions/models/ Push to the Swamp registry: swamp extension push ./manifest.yaml Create a *Claw skill that pulls the extension and uses it: swamp extension pull This creates a two-layer sharing model: extensions for model types (Swamp registry), skills for agent workflows (ClawHub). ### Model a REST API "Model the JSONPlaceholder API so I can list and create posts" swamp model create command/shell jsonplaceholder Edit to add methods: list-posts (GET /posts), create-post (POST /posts) swamp model method run jsonplaceholder list-posts Verify output, iterate ### Set Up a Weather Integration "Create a weather model that fetches forecasts by city" swamp model create command/shell weather Edit to add a forecast method using curl -s "wttr.in/{city}?format=j1" Test with swamp model method run weather forecast ### Chain APIs in a Workflow "Create a workflow that fetches GitHub issues, summarizes them, and posts to Slack" Create models: github-issues, slack-webhook Create workflow: swamp workflow create issue-digest Define jobs with dependencies: fetch issues -> format summary -> post to Slack swamp workflow run issue-digest ### Graduate to a *Claw Skill "Turn my working weather model into a skill others can install" swamp model get weather --json to export Create a new SKILL.md wrapping the weather model commands clawhub publish ./weather-skill ### Share a Custom Extension "Publish my custom Stripe model type so the team can use it" swamp auth login to authenticate with swamp-club swamp extension push ./stripe-models/manifest.yaml --dry-run to verify swamp extension push ./stripe-models/manifest.yaml to publish Team members install: swamp extension pull stripe-models ### Important Notes Always use --json flag when you need to parse Swamp output programmatically Use swamp model validate before running to catch definition errors early Store all credentials in vaults, never in model definitions Use -v (verbose) flag when debugging unexpected behavior All Swamp operations run locally — credentials stay on the user's machine Authenticate with swamp auth login before pushing extensions to the registry Use swamp extension push --dry-run to verify an extension before publishing Use swamp auth whoami to confirm your identity before registry operations ## Trust - Source: tencent - Verification: Indexed source record - Publisher: umag - Version: 0.2.0 ## Source health - Status: healthy - Item download looks usable. - Yavira can redirect you to the upstream package for this item. - Health scope: item - Reason: direct_download_ok - Checked at: 2026-05-12T18:03:23.184Z - Expires at: 2026-05-19T18:03:23.184Z - Recommended action: Download for OpenClaw ## Links - [Detail page](https://openagent3.xyz/skills/swamp) - [Send to Agent page](https://openagent3.xyz/skills/swamp/agent) - [JSON manifest](https://openagent3.xyz/skills/swamp/agent.json) - [Markdown brief](https://openagent3.xyz/skills/swamp/agent.md) - [Download page](https://openagent3.xyz/downloads/swamp)