Content
66%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 presents a clear confirm-scope → read-queue → present-results workflow with explicit validation checkpoints and strong failure-mode handling, but it leans on a helper script that is absent from the bundle and restates its authorization caveats in every section. Commands are concrete in form, yet the primary execution path is not runnable as shipped due to the missing and inconsistently-pathed script.
Suggestions
Bundle the helper script (e.g. under scripts/) or make its path resolvable, and align the "../../../scripts/github-review-requests.mjs" location with the "scripts/..." invocation run from the repository root.
Consolidate the repeated authorization/scoping caveats into the "Confirm the scope" section instead of restating them in each section.
Drop the checkout-specific hardcode ("This checkout belongs to BigSimmo/Communication") or generalize it into a clearly-labeled example.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient — it explains no concepts Claude already knows — but the authorization/scoping caveats repeat across every section ("Never assume an organization from this copied skill", "do not switch to another organization", "Never fall back to account-wide notifications/search or a different owner/team"), and the hardcoded "This checkout belongs to `BigSimmo/Communication`" is environment-specific noise. Not 4: the repetition is more than minor over-explanation; not 2: there is no padded conceptual exposition. | 3 / 5 |
Actionability | Concrete commands are given ("node scripts/github-review-requests.mjs --repo ExampleOrg/example-repo --team ExampleOrg/example-team", "git remote get-url origin", "gh pr list --repo github.com/OWNER/REPO --state open --search 'team-review-requested:OWNER/TEAM'"), but the primary helper is referenced at "../../../scripts/github-review-requests.mjs" and no scripts/ directory exists in the bundle, and the invocation path ("scripts/..." from the repository root) is inconsistent with that location — so the main path is not executable as shipped. Not 4: a missing helper script is a key gap, not a minor one; not 2: the commands themselves are specific and complete in form, not high-level hints. | 3 / 5 |
Workflow Clarity | A clear sequence (Confirm the scope → Read the scoped queue → Present results and limits) with explicit validation checkpoints: verify the remote host via `git remote get-url origin` before deriving OWNER/REPO, "Stop for GitHub Enterprise", the helper "validates the operands" and "checks that the PR is open, unmerged and still requests that team/reviewer", and "Report zero only after a successful lookup; unavailable tooling, denied access and a failed request are not an empty queue" — an explicit failure-recovery distinction. The skill is read-only, so the destructive/batch cap does not apply. Not 4: essentially all checkpoints are present, including failure-mode handling. | 5 / 5 |
Progressive Disclosure | The body is short (~40 lines) and organized into three well-signed sections, fitting the simple-skill pattern where good organization alone can score highly; however, it depends on a helper whose referenced path ("../../../scripts/github-review-requests.mjs") does not resolve — no references/, scripts/, or assets/ directories exist in the bundle. Not 5: the one external reference is dangling; not 3: nothing that belongs in a separate file is inlined and the structure is otherwise clear. | 4 / 5 |
Total | 15 / 20 Passed |