| name | workflows-mcp |
| description | Use when onboarding and operating workflows-mcp from an MCP client in any project, including install, client configuration, memory setup, project_onboard/project_sync usage, and workflow authoring/execution. |
workflows-mcp-user
User-facing skill for adopting workflows-mcp in your own project.
When to use
Use this skill when you want to:
- install and run
workflows-mcp from an MCP client,
- configure custom workflow directories and secrets,
- enable PostgreSQL-backed memory,
- use
project_onboard / project_sync checkpoint flows,
- create, validate, and execute workflows.
What you get
- Fast install paths (
uvx or pip)
- Copy/paste MCP client config snippets
- Memory setup checklist (PostgreSQL + env vars)
- Practical memory operation patterns (
query, ingest, lifecycle, graph)
- Workflow authoring + validation + execution loop
- Best practices, pitfalls, and troubleshooting steps
Start here
- Read
README.md in this skill package.
- Copy config from
snippets/ (uvx or pip).
- Follow docs in order:
docs/quickstart.md
docs/memory-postgres.md
docs/workflow-authoring.md
Ground rules
- Commands and payloads are aligned to current public docs and tool contracts.
- Use
reload_workflows after editing workflow YAML files.
- Restart your MCP client after changing server env/config.
- Keep secrets in
WORKFLOW_SECRET_* env vars (never inline in workflow files).