Content
71%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 pipeline skill with concrete commands, templates, decision tables, and a complete worked example. Its weaknesses are the absence of any external reference files (forcing all lookup material inline in a long SKILL.md) and the lack of post-generation validation or error-recovery steps. Referenced script paths also don't resolve within this bundle.
Suggestions
Move the Platform Specs table, Camera Language Reference, and FPV/drone prompt-enrichment blocks into a references/ file (e.g. camera-and-platform-specs.md) and link to them from the workflow steps.
Add an explicit validation/error-recovery step after Step 5 — e.g. what to check in the generated video (duration, aspect ratio, brand consistency) and what to do if the script fails.
Document the --duration flag in the run-social-video.sh invocation block and verify the referenced script paths (core/media/generate-image.sh) are reachable from the skill's location.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and efficient — tables for brand files, modes, image models, and platform specs, plus a Director Brief template, with almost no explanation of concepts Claude already knows. It is not 5 because 'Prompt Quality Checklist' and 'Common Mistakes to Avoid' partially repeat each other (sound direction, vague adjectives, generic prompts), which could be trimmed. | 4 / 5 |
Actionability | Guidance is mostly executable: concrete script invocations with flags, mode-selection tables, image-model recommendations by use case, and a full worked example with prompt text. It is not 5 because the referenced scripts (core/media/generate-image.sh, library/social/social-media-video/scripts/run-social-video.sh) do not exist in this bundle, and the example uses an undocumented '--duration 11' flag — minor gaps that keep it from copy-paste-ready throughout. | 4 / 5 |
Workflow Clarity | The five-step Agent Execution Protocol is clearly sequenced with input-gap handling ('If a file is missing, proceed with what's available and note the gap') and a pre-flight Prompt Quality Checklist. It is not 5 because there is no validation or error-recovery after generation — no instruction on what to do if a script fails or how to verify the output video meets platform specs. | 4 / 5 |
Progressive Disclosure | Sections are well-organized, but no bundle files exist — the ~100 lines of platform specs and camera-language lookup tables, plus the FPV/drone prompt-enrichment blocks, are inlined in a 270-line SKILL.md where they belong in reference files. It is not 4 because content that clearly should be separate (lookup/reference material) is inline; not 2 because headers and structure make it navigable. | 3 / 5 |
Total | 15 / 20 Passed |