Skip to content

docs: apply the operational data positioning to the README, manifests, skills and evals - #20

Open
HasanAlpCaferoglu wants to merge 1 commit into
mainfrom
docs/operational-positioning
Open

HasanAlpCaferoglu wants to merge 1 commit into
mainfrom
docs/operational-positioning

Conversation

@HasanAlpCaferoglu

Copy link
Copy Markdown
Collaborator

What does this change?

Applies the operational data positioning from DataSQRL/sqrl#2608 (POSITIONING.md, the blog post and the comparison page) to the plugin. The plugin version is now 0.4.1.

  • Naming: "DataSQRL Code Agent" is now "DataSQRL agent" in all files. The open-source edition is the advanced agent (ghcr.io/datasqrl/adv-agent). The pro edition is the pro agent (ghcr.io/datasqrl/pro-agent), which is not open source. The launcher image names do not change.
  • README: new title "DataSQRL Plugin for Coding Agents", the tagline, the 25-word definition, a "When to use it" section, an operational example prompt, the reason for verification, and a table of the two agents in Requirements. Each paragraph is now on one line.
  • Manifests: the new description and seven more keywords in all four manifests. The Codex manifest has new shortDescription, longDescription and defaultPrompt texts. The marketplace description now describes the catalog.
  • start skill description: adds the triggers Kafka, CDC, data product and operational data pipeline. All old triggers stay.
  • start skill, "Using DataSQRL": opens with the boilerplate, defines operational data, and names SQRL. It lists the six fit signals and a rule to decide the fit by where the work runs and who consumes the result. It adds the warehouse fit (CDC into Iceberg, with views for each query engine) and the non-fit for scheduled models in a warehouse (dbt or SQLMesh). It lists the tools that DataSQRL works with and does not replace.
  • Evals: two new start cases, dbt-models-not-a-fit and warehouse-landing-is-a-fit.
  • Repository metadata: the About line and ten topics are already set on GitHub.

Area

  • Skills
  • Launcher (scripts/datasqrl-agent.sh)
  • Cloud client (scripts/datasqrl-cloud.sh)
  • Manifests
  • CI
  • Docs

scripts/datasqrl-agent.sh changes only in one comment and one error message, for the rename.

How did you test it?

  • .github/check-plugin.sh passes with version 0.4.1. shellcheck was not installed locally, so only CI runs that lint step.
  • pytest tests: 174 passed. This run was before the README reflow, the marketplace description change and the version change. The tests read none of those files.
  • Evals in tests/claude-plugin-evals/skill-specific-tests, 3 runs each: all 13 cases scored 1.00, with the two new cases. The four start cases scored 1.00 again after the last change to the start skill.
  • Stub launcher eval patch-small-change-runs-patch: 0.92 against the 0.8 threshold. One of 3 runs did not give the run name in its reply.
  • Not tested: an install of the plugin in Codex, Cursor or Copilot.

Checklist

  • The PR title starts with a type such as feat:, fix: or docs:
  • .github/check-plugin.sh passes
  • Skills contain no SQRL and no HTML comments
  • A launcher change that needs a new agent image option waits until that image is published
  • README or DATA_HANDLING.md is updated if behavior or data handling changed

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant