Content
71%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.
A well-structured, highly actionable API catalog: every endpoint has an executable example and the setup flow includes a credential fallback. Its main weaknesses are redundancy (duplicated endpoint descriptions and repeated curl boilerplate) and the absence of progressive disclosure — the full endpoint reference lives inline in SKILL.md instead of a separate file.
Suggestions
Move the per-endpoint parameter lists and curl examples into a references/ file (e.g. references/endpoints.md), keeping SKILL.md to setup, a capability summary, and 2-3 representative examples.
Remove the duplication between the "## Capabilities" bullet list and the per-endpoint descriptions in "## Usage" — state each capability once.
Add brief error-handling guidance (e.g., what to do on 401/429 responses or empty results) to close the workflow-clarity gap.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body contains no concept-padding, but each endpoint's one-line description is stated twice ("## Capabilities" list plus each "### Usage" subsection) and the identical curl boilerplate (URL, headers) repeats verbatim across 20 endpoint blocks, which could be tightened considerably. It is above the verbose anchor of 2 because nearly all content is operational data Claude does not already know. | 3 / 5 |
Actionability | Every one of the 20 endpoints ships a complete, executable curl command with realistic values (e.g. '-d '{"api":"tomba","path":"v1/email-finder","query":{"domain":"stripe.com","first_name":"John","last_name":"Doe"}}''), plus a concrete credential bootstrap with a fallback login command — copy-paste ready and covering the common cases. | 5 / 5 |
Workflow Clarity | The sequence is clear — credential setup (with an explicit checkpoint: "If ~/.gooseworks/credentials.json does not exist, tell the user to run: npx gooseworks login") then a single unambiguous API call per task, with no destructive or batch operations to cap the score. It falls short of 5 only because there is no guidance on handling API errors or unexpected responses. | 4 / 5 |
Progressive Disclosure | Roughly 250 lines of per-endpoint API reference are inlined in SKILL.md with no separate reference files (no references/, scripts/, or assets/ exist in the bundle), matching the anchor of content that should be separate being inline. Section headers keep it navigable and the "Discover More" section points at live endpoint-discovery endpoints, so it is above the minimal-structure anchor of 2 but below the good-structure anchor of 4, which expects the bulk reference moved out of the overview. | 3 / 5 |
Total | 15 / 20 Passed |