Content
72%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 CLI reference skill with excellent actionability — every command is concrete and copy-paste ready. The main weaknesses are the lack of validation/verification steps for destructive operations (delete-slide, replace-text, batch-update) and some reference material that could be better organized into separate files. The skill reads more like a CLI man page than a workflow guide, which limits its workflow clarity score.
Suggestions
Add verification steps after write operations, e.g., 'After replace-text, run get-text to confirm changes were applied correctly' and 'Before delete-slide, use get-metadata to verify the slide ID'
Add a feedback loop for batch-update: validate the JSON input, check the response for errors, and verify results with get-text or get-metadata
Consider moving the Slide Layouts list and Output Format examples into a separate REFERENCE.md file to keep the main skill focused on workflows
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Generally efficient with concrete commands and examples. The 'Presentation ID Format' and 'Output Format' sections add useful reference value. Minor verbosity in the 'Slide Layouts' list (Claude could look these up) and the 'Token Management' section which explains system keyring details Claude doesn't need. | 4 / 5 |
Actionability | Fully executable, copy-paste ready commands covering all common operations. Every command includes concrete examples with realistic arguments, flags, and variations. The batch-update example even shows the JSON request format. | 5 / 5 |
Workflow Clarity | The batch-update command is a destructive/batch operation with no validation or verification steps. There's no guidance on checking results after write operations (e.g., verifying replace-text worked correctly, confirming slide deletion). The delete-slide command references get-metadata for finding IDs but doesn't show a complete workflow with verification. This caps the score at 3 per the rubric's feedback loop requirement for batch/destructive operations. | 3 / 5 |
Progressive Disclosure | Content is reasonably structured with clear section headers, but the output format examples and slide layouts list could be in separate reference files. With no bundle files provided, there's no evidence of content being split across files. The skill is moderately long (~100 lines) and could benefit from separating reference material from core usage instructions. | 3 / 5 |
Total | 15 / 20 Passed |