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 tightly written, well-sectioned procedural skill: concise, domain-specific, and honest about boundaries, with the bulk of detail appropriately deferred to external references. The main weaknesses are unnamed resources (the `ads` operating contract, the shared reference files, the common JSON contracts) that force discovery rather than direct navigation, and the absence of an explicit error-recovery feedback loop in the procedure.
Suggestions
Name or link the exact files for step 1 ("the main `ads` operating contract") and step 3's shared references (e.g., `ads/references/measurement.md`, `ads/references/benchmarks.md`), the same way `ads/references/tiktok-audit.md` is named, so navigation does not depend on exploration.
Identify the concrete JSON contract or schema file to use in step 7 and the Output section ("the common JSON contracts") so schema-valid findings can be verified directly rather than inferred.
Add a short error-recovery step to the Procedure (e.g., what to do when required inputs are missing or exports conflict) to complement the existing validation gates and lift workflow clarity.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~30-line body is lean with zero concept over-explanation — no preamble about what TikTok Ads is or why auditing matters; it assumes Claude's competence and every step earns its place. Even the parenthetical path-resolution instruction in step 3 is operational information, not padding, so it fits the 5 anchor rather than the 4 anchor ('minor instances of over-explanation'). | 5 / 5 |
Actionability | Mostly executable guidance: step 2 enumerates the exact inputs to collect ("objective, conversion definition, account and campaign age, geography, date window, timezone, currency, spend, targets"), and step 3 names `ads/references/tiktok-audit.md` with path-resolution instructions. However, step 1's "the main `ads` operating contract and thinking framework" and step 7/Output's "common JSON contracts" are never named or located, leaving minor gaps — matching the 4 anchor rather than 5 (fully executable) or 3 (incomplete with missing key details for most steps). | 4 / 5 |
Workflow Clarity | The 8-step Procedure is a clear, well-ordered sequence with several validation gates ("Do not apply a benchmark without checking objective, geography, methodology...", "Return schema-valid findings", "Keep every account change as a draft until the main mutation gate passes"). It falls short of the 5 anchor because there is no explicit validate-then-retry feedback loop or error-recovery step (e.g., what to do when inputs in step 2 are unavailable beyond listing them as missing); this is a read-only audit, so the destructive-operation cap does not apply. | 4 / 5 |
Progressive Disclosure | Good structure: a short overview body with three clear sections, and the detailed material is correctly pushed out to a clearly signaled one-level-deep reference (`ads/references/tiktok-audit.md`) with explicit resolution paths. It stops short of the 5 anchor because the "relevant shared measurement, benchmark, creative, automation, policy, and scoring references" are named only by category, not by file, making navigation slightly less crisp; no bundle files exist in this skill directory, so the structure relies on the external plugin layout. | 4 / 5 |
Total | 17 / 20 Passed |