Content
82%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is highly actionable — every section carries executable commands or complete code — with a well-sequenced test workflow and real bundle-based progressive disclosure. It loses points for duplicated setup commands across sections, a missing readiness/validation checkpoint before applying metadata, and inline content that belongs in the references directory.
Suggestions
Consolidate the repeated 'docker compose up' / 'hasura metadata apply' commands into a single place (e.g. the Running section) and reference it from How to use and Authoring to remove duplication.
Add an explicit readiness check before applying metadata (e.g. curl-wait loop on /healthz or the v1/metadata endpoint), since docker 'depends_on' does not wait for the engine to accept requests — this closes the workflow's main validation gap.
Move the CI workflow YAML and the inline permission-test code into the existing references/ directory (or a new references/ci-integration.md) so SKILL.md stays a lean overview with one-level-deep links.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean and command/code-driven with no basic-concept padding, but 'docker compose up', 'hasura metadata apply', and the admin-secret handling repeat across 'How to use', 'Authoring', and 'Running', and the References section restates links already cited inline. | 4 / 5 |
Actionability | Fully executable throughout: a complete docker-compose.test.yml with required-variable guards, curl and hasura CLI commands with headers, copy-paste-ready httpx/pytest code with real assertions, and a complete CI workflow YAML — covering the common cases. | 5 / 5 |
Workflow Clarity | A clear 7-step sequence (stack up → metadata apply → per-test seeding → role-override queries → assert → matrix + JWT smoke test → teardown with -v and metadata reset) with explicit assertion checkpoints and an anti-patterns table; minor gaps include no engine-readiness check before hitting :8080 and no explicit failure-recovery loop. | 4 / 5 |
Progressive Disclosure | Good structure with a single one-level reference (references/permission-matrix-tests.md, verified present) clearly signaled and a well-organized References section; however the ~30-line CI YAML and the inline permission-test code (which overlaps the reference file's content) could be pushed into references to keep SKILL.md a leaner overview. | 4 / 5 |
Total | 17 / 20 Passed |