Content
67%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 process skill: clear phased workflow with hard caps, concrete commands, worked output templates, and explicit anti-patterns. Its main weaknesses are a confusing, irrelevant opening paragraph, duplicated cap statements, and placeholder/slightly-broken search examples; validation of generated tests is delegated rather than inlined.
Suggestions
Delete the 'Artifact consistency before implementation' section — its seat/escalation jargon ('a clean pass costs no seat', 'at most one independent seat') has no role in the coverage workflow and wastes context.
Deduplicate the limits: the 'Red Flags' table repeats the 30-path, 20-test, and 2-minute caps already stated in 'Caps and Limits' — keep only the unique anti-pattern rows there.
Fix the search examples in Phase 2: replace the placeholder "grep -rl \"[function_name]\" tests/" (brackets form a regex character class) with a concrete, executable command, and show one fully substituted example.
Inline a run-tests → fix-failures → re-run loop at the end of Phase 4 instead of only delegating verification to skill-verification-gate, so generated tests are validated within the skill's own workflow.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Most of the body is tight and prescriptive (commands, tables, output templates), but there is more than minor waste: the "Artifact consistency before implementation" paragraph is impenetrable policy jargon ("A single artifact skips analysis; a clean pass costs no seat", "Escalation uses at most one independent seat") unrelated to the coverage workflow, and the 30-path / 20-test / 2-minute caps from "Caps and Limits" are restated nearly verbatim in three Red Flags table rows. This fits anchor 3 (mostly efficient, some unnecessary explanation, could be tightened) better than anchor 4's 'minor instances that could be trimmed'. | 3 / 5 |
Actionability | Phase 1 supplies copy-paste git commands and Phase 2 gives concrete find/grep searches, structured inventory/coverage-map templates with worked example rows, a star-rating rubric, and a conventions-detection checklist — mostly executable guidance. It misses anchor 5 because several search commands are unsubstituted placeholders and one is subtly wrong as written ("grep -rl \"[function_name]\" tests/" — the bracketed token is a regex character class, not a literal placeholder), which are minor gaps in executability. | 4 / 5 |
Workflow Clarity | The sequence is explicit and well-ordered (TRACE → MAP → SCORE → DIAGRAM → GENERATE → REPORT, with per-phase steps, caps, and priority rules), and the batch test-generation phase does name a verification checkpoint ("use skill-verification-gate to run the test suite and confirm the new tests pass"), so the missing-validation cap of 3 does not apply. It stops short of anchor 5 because validation is delegated to another skill rather than inlined as a run → fix-failures → re-run feedback loop, and no checkpoint verifies the codepath inventory or coverage map itself. | 4 / 5 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are all absent) and the skill is a single ~250-line file with clear section headers, an overview, and a Quick Reference that make navigation easy — most content is appropriately placed inline. It does not reach anchor 5 because the file is on the larger side and self-contained material that could live one level deep (e.g., the example inventory/coverage-map tables or the integration guide) is inlined, which is a minor organization gap rather than an ideal split. | 4 / 5 |
Total | 15 / 20 Passed |