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-organized, tool-specific reference with strong executable examples and explicit error-inspection guidance for the batch operation. The notable defect is progressive disclosure: the body's key external reference (references/cli-reference.md) and the sibling-skill API references do not exist in this bundle, and the Python example has a top-level await that won't run as written.
Suggestions
Fix the dangling reference: either include references/cli-reference.md in the bundle or point to its actual location, and clarify where the python/nodejs/rust API references live relative to this skill.
Make the Python example executable by wrapping the await in an async function with asyncio.run(), so the snippet is copy-paste ready.
Trim soft filler sentences ('A single shared config does not always fit.') and the intro line that restates the description to tighten token efficiency.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with tool-specific knowledge (defaults, concurrency limits, layout caps, path-key semantics) and assumes Claude's competence. Minor trimmable filler keeps it below anchor 5: the intro restates the description, and sentences like 'A single shared config does not always fit.' add little. Well above anchor 3 — no explanations of concepts Claude already knows. | 4 / 5 |
Actionability | Nearly all guidance is copy-paste executable: concrete bash commands, a jq results-inspection pipeline, a complete JSON file-configs example, and a TOML config. Falls short of anchor 5 because the Python example uses top-level `await extract_batch(inputs, config)` outside any async function or runner, so it is not executable as written. | 4 / 5 |
Workflow Clarity | Sections are logically ordered (usage, parallelism, overrides, output, error recovery, config, programmatic access) and the batch operation includes an explicit verification step — 'Inspect results for partial content and surfaced errors rather than relying on the process exit code alone' — so the batch-cap at 3 does not apply. Missing an explicit validate→fix→retry loop keeps it at anchor 4 rather than 5. | 4 / 5 |
Progressive Disclosure | References are clearly signaled and one level deep, but scored against the actual bundle: no `references/` directory exists, so 'See `references/cli-reference.md` for the full `batch` flag set' is a dangling pointer, and the API references live in a 'sibling `xberg` skill' that is not part of this bundle. The primary pointer for the full flag set is broken, which is more than the minor organization gaps of anchor 4; better than anchor 2 since the inline structure is well organized. | 3 / 5 |
Total | 15 / 20 Passed |