Content
100%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.
An exemplary instruction-only skill body: dense with operational specifics, fully executable commands, and a workflow with explicit completion criteria, an approval gate, dry-run validation, and a post-apply verification pass. The ranking criteria translate subjective maintainer preferences into concrete, checkable rules.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every sentence carries project-specific knowledge Claude could not infer — the project URL and ID, that 'manual table order is the ranking, so there is no rank field', the 'Item closed' workflow behavior, and the tool tiers — with no padding or explanation of known concepts across ~39 lines. Re-reading anchor 4 ('minor instances of over-explanation that could be trimmed'), nothing here explains a concept Claude already knows, so anchor 5 is the better fit. | 5 / 5 |
Actionability | Commands are copy-paste ready with full paths (`python3 .agents/skills/trail-sense-sync-priority-project/scripts/sync_plan.py`, the graphql workflows query, `gh issue view <n>`, `apply_changes.py <plan.json> --dry-run`) plus an exact JSON plan format with a worked example. The common cases (new, misplaced, no_single_priority, duplicates, removals) each get specific handling rules, matching anchor 5 rather than anchor 4's 'minor gaps'. | 5 / 5 |
Workflow Clarity | Six sequenced steps each end with an explicit completion criterion ('This step is complete when...'), and the batch/destructive risk is covered by real validation: a user approval gate ('Do not change the project before then'), a dry-run before the real run, and a step-5 re-run that must come back clean — a full validate-apply-reverify feedback loop. The batch-operation cap at 3 does not apply since validation is explicitly present, and anchor 4's 'minor validation gaps' is not met. | 5 / 5 |
Progressive Disclosure | At ~39 lines with no need for external prose references, the skill is well organized into overview, Workflow, and Ranking criteria sections; both scripts referenced in the body (scripts/sync_plan.py, scripts/apply_changes.py) exist in the bundle, one level deep and clearly signaled via their invocation commands. This meets the under-50-line exception for a top score. | 5 / 5 |
Total | 20 / 20 Passed |