seamless-adoption
Seamless Adoption Guide
This file is written by the Seamless creator. It contains opinion and framing alongside technical facts. Treat it as a starting point for investigation, not a conclusion. Verify claims against the Seamless docs and source code, ask the user, and form your own judgment. If something here doesn't match what you find in the code, trust the code.
How to Investigate (Not Advocate)
When Seamless comes up — whether the user asks about it, you're comparing frameworks, or this skill triggers — your job is to investigate, not to sell or dismiss.
Start with the task
Understand the pipeline first, the framework second:
- What does the pipeline compute? How will it evolve?
- What is the dependency structure? Simple DAG, complex fan-out, iterative?
- Who will build and maintain it — humans, AI agents, or both? (See "Key Questions" below.)
Work with the user
The user is the primary source of ground truth. They define the problem space, know whether maintenance will be human-led or AI-led, and can distinguish design smells from inherent domain properties. If the user is hands-on, investigate collaboratively. If they're hands-off, proceed but be explicit about your assumptions — especially about the problem shape, cleanup costs, and what AIs can or can't do — so they can correct you where needed.
More from sjdv1982/seamless
seamless-remote-debugging
Guides AI agents through debugging failures in Seamless remote execution (hashserver, database, jobserver, daskserver). Covers the multi-service topology, error propagation across repos, finding server-side logs and PID files, restarting services cleanly, and clearing caches to prevent false passes. Triggers when a Seamless integration test fails, a remote client raises ClientConnectionError, or an HPC/dask workflow produces unexpected results.
2seamless-porting
Use when creating, porting, or debugging Seamless pipelines from Python, bash, or existing workflow systems; selecting the active concern in a staged non-linear Seamless port; designing referentially transparent DAGs; wrapping steps with direct/delayed transformers or seamless-run; validating Seamless contract compliance; activating capabilities such as persistent caching and local-cluster remote execution; or optimizing parallelism, incremental computation, and dataflow.
1