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.
A dense, highly actionable CLI skill: every section is driven by executable commands or complete YAML, risky operations carry explicit confirmation guidance, and detail is pushed to clearly signaled one-level-deep reference files. The deductions are minor — a few inline sections that duplicate reference-file detail, parity verification placed in the mistakes table instead of the workflow, and reference files absent from the evaluated bundle.
Suggestions
Trim inline detail that the reference files already own — e.g., move the full DEPOT_ACTOR_*/DEPOT_TRIGGERING_ACTOR_* environment-variable enumeration and the dispatch actor-context paragraph into a reference file, keeping only the depot.actor vs depot.triggering_actor distinction inline.
Fold parity verification into the Getting Started sequence as an explicit numbered step (e.g., "4. Run a workflow in both engines and compare the reported checks") instead of leaving it implicit in the Common Mistakes table.
Ship the referenced files (references/migration.md, references/secrets-and-variables.md, references/oidc.md, references/runs-and-debugging.md, references/github-actions-compatibility.md, references/posthog-check-run-semantics.md) in the skill bundle so the progressive-disclosure links resolve.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with product-specific facts Claude cannot know (variant selectors, `snapshot:` syntax, retry/backoff semantics, `--forge` detection rules) and contains no general-concept padding, but a few stretches spend tokens on detail the reference files already own — the eight-variable "DEPOT_ACTOR_KIND ... DEPOT_TRIGGERING_ACTOR_NAME" enumeration and the 13-bullet compatibility index. Fits anchor 4 (efficient with minor trimmable over-explanation); not 5 because those sections exceed what the overview needs, and not 3 because nothing explains concepts Claude already knows. | 4 / 5 |
Actionability | Copy-paste-ready bash and YAML throughout: "depot ci migrate", "printf '%s' "$NPM_TOKEN" | depot ci secrets add SECRET_NAME", "depot ci run --workflow .depot/workflows/ci.yml --job build --ssh-after-step 3", complete `snapshot:`/`parallel:`/`retry:` YAML blocks, and a Common Mistakes table with concrete fixes. Common cases are covered in full; not anchor 4 because nothing is pseudocode and no key flags are missing. | 5 / 5 |
Workflow Clarity | Getting Started gives a clear install → migrate → import-secrets sequence, the migration wizard validates compatibility as it goes, and risky credential work has an explicit checkpoint ("Before running credential-affecting commands, confirm scope (org, repo, workflow) and intended target"), with `depot ci diagnose` as an error-recovery loop. Fits anchor 4; not 5 because post-migration parity verification ("Keep them during transition to verify Depot CI parity") lives in the Common Mistakes table rather than as an explicit step in the workflow itself. | 4 / 5 |
Progressive Disclosure | Signaling is exemplary and one level deep — each section states exactly when to load its reference ("For migration detail ... read `references/migration.md`", "you don't need to load `references/github-actions-compatibility.md`" for routine tasks). However, no `references/` directory exists in the bundle, so the six referenced files (migration.md, secrets-and-variables.md, oidc.md, runs-and-debugging.md, github-actions-compatibility.md, posthog-check-run-semantics.md) cannot be verified as real; the structure merits 5 but the unverifiable pointers hold it to anchor 4's "minor organization gaps". | 4 / 5 |
Total | 17 / 20 Passed |