Content
53%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.
The body is a well-organized catalog stub that is honest about its upstream-maintenance model and gives concrete install/prerequisite commands. However, it instructs almost nothing about executing the skill's own workflow — the pipeline is narrated rather than prescribed — and it spends tokens on catalog and install-status meta-commentary. It functions acceptably as a pointer to the real skill but not as a standalone SKILL.md.
Suggestions
Trim the catalog-maintenance meta-commentary (the stub blockquote and the skills-guard status aside) to a one-line pointer, keeping only the install command and docs URL.
Prescribe the actual workflow as numbered steps (read project → pick tone → storyboard → composition brief → render → verify output) with a validation checkpoint that brag-output/brag.mp4 and the poster were produced.
Move the time-sensitive install-status details ("hyperframes-creative scores dangerous, which --force cannot override") into the upstream README or a separate status file, since they will go stale.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The prerequisites section is dense and useful ("Node.js 22+, FFmpeg on PATH", "npx hyperframes doctor", GITHUB_TOKEN guidance), but the 8-line "Catalog stub" blockquote and the upstream-maintenance meta-commentary ("quarantined and scanned like any hub install") explain catalog plumbing rather than the task. Mostly efficient with some unnecessary explanation — anchor 3. | 3 / 5 |
Actionability | There are concrete, runnable commands ("hermes skills install official/creative/brag", "npx hyperframes doctor") and a fully enumerated options list ("--tone", "--format landscape|vertical|square", "--voice"), but the core pipeline ("commits to an angle and a tone", "hands a composition brief to Hyperframes") is only described, never instructed. Some concrete guidance but incomplete — anchor 3. | 3 / 5 |
Workflow Clarity | The pipeline sequence is implicit in prose ("reads the project... storyboards a 15–25 second video, hands a composition brief to Hyperframes, and renders brag-output/brag.mp4") with no numbered steps, checkpoints, or error-recovery guidance. This is a multi-step render workflow, so the simple-skill exception does not apply; anchor 3 ("steps listed but validation gaps... sequence present but checkpoints missing or implicit"). | 3 / 5 |
Progressive Disclosure | At ~45 lines with well-organized sections, the body clearly signals one-level-deep external pointers ("Full documentation: https://github.com/latent-spaces/brag#readme", upstream "skills/brag/" with its step references and scripts) and explicitly explains that this directory holds only catalog metadata. Not 5 because the referenced bundle files (scripts/analyze_music_cues.py, tone presets, music library) are not present locally and navigation depends entirely on the external repo. | 4 / 5 |
Total | 13 / 20 Passed |