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, actionable skill body with a clear sequenced workflow, explicit validation and error-handling steps, and executable commands anchored by a real script. Its weaknesses are redundancy (scope-guard rules repeated across four sections plus overlapping Stress-Case/Response-Template blocks) and an undocumented, misused --style flag in one usage example.
Suggestions
Consolidate the scope-guard messaging repeated across When to Use, Scope Boundaries, Input Validation, and Error Handling into a single section, and merge Stress-Case Rules into the Response Template section.
Document the --style flag (academic/conference/lightning) in the Parameters table and fix the invalid "--style minimal" in the third usage example to a valid value such as "--style conference" or use "--template minimal".
Add an explicit validate-and-retry loop for the PDF validation step (e.g., after the parse error, accept a text/abstract fallback and re-run extraction) to strengthen the workflow's feedback structure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient — a compact parameter table, short usage block, no explanation of concepts Claude already knows — but scope-guard messaging is repeated across four sections: "Validate that the request is for presentation generation from existing content, not original research writing" (Workflow), "Input Validation", "Scope Boundaries" ("it does not write original research"), and "Error Handling". "Stress-Case Rules" and "Response Template" also overlap heavily (both mandate Assumptions/Risks blocks). This redundancy could be consolidated into one section, matching 'mostly efficient but could be tightened'. | 3 / 5 |
Actionability | Guidance is largely executable: a Quick Check ("python -m py_compile scripts/main.py", "python scripts/main.py --help"), three concrete usage commands, and a complete-looking parameter table. The gaps: the third example uses "--style minimal" yet `--style` is absent from the Parameters table entirely, and the script's actual choices (academic/conference/lightning) exclude "minimal", so that example fails as written. That internal inconsistency keeps it below the fully copy-paste-ready 5. | 4 / 5 |
Workflow Clarity | The 6-step Workflow is clearly sequenced with explicit validation checkpoints — "PDF Validation: ... If the PDF is encrypted, image-only, or corrupt, emit a specific error" and step 3's request-scope validation — plus an Error Handling section with recovery guidance ("report the failure point, summarize what still can be completed safely, and provide a manual fallback"). It is not a 5 because there is no validate-then-retry feedback loop after a failure (e.g., re-running after fixing a bad PDF) and the input-validation step is duplicated rather than chained as an explicit checkpoint. | 4 / 5 |
Progressive Disclosure | Structure is good: clearly titled sections (Quick Check, When to Use, Workflow, Usage, Parameters, Output, Scope Boundaries, Error Handling) and a real, working reference to scripts/main.py one level deep — the referenced bundle file exists and its flags match the documented interface (except the undocumented --style). The body is ~100 lines, over the under-50-lines threshold where organization alone earns a 5, and the Stress-Case/Response-Template sections could live in a reference file, so it sits at 'good structure; most content appropriately placed'. | 4 / 5 |
Total | 15 / 20 Passed |