Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Reviews and iterates on social media copy in album promo/ files. Use after populating promo templates and before release to polish platform-specific posts.
.claude/skills/bitwize-music-studio-promo-reviewer/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 149% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 184% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 125% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 96% | 0% |
Interactive review and polish of social media copy in album promo/ files. Walk through each post, approve or revise, and write polished results back to the file.
After promo templates are populated with platform-specific copy, this skill provides structured review before release. Each post is presented with character counts, hashtag counts, and platform limit compliance. The user chooses actions (approve, revise, shorten, punch up, etc.) and polished copy is written back.
/bitwize-music:promo-writer generates initial copyPromo Videos (optional) → [Promo Writer] (or manual) → **[Promo Review]** → ReleaseBetween populating promo templates and release-director.
Resolve the album from arguments:
Use MCP find_album with the album name from $ARGUMENTS. If no album specified, check get_session for last album context.
Locate promo directory:
{content_root}/artists/{artist}/albums/{genre}/{album}/promo/Check which promo files exist and have content:
Glob for promo/*.md in the album directory. For each file, check if it contains populated content (not just template placeholders). A file is "populated" if it contains text beyond the template markers — look for content inside code blocks that isn't [placeholder text].
Report status:
## Promo Copy Status
| Platform | File | Status |
|----------|------|--------|
| Campaign | campaign.md | Populated |
| Twitter/X | twitter.md | Populated |
| Instagram | instagram.md | Template only |
| TikTok | tiktok.md | Not found |
| Facebook | facebook.md | Populated |
| YouTube | youtube.md | Template only |
3 of 6 platforms have copy ready for review.If no promo files are populated:
No promo copy found to review.
Options:
1. Generate promo copy: /bitwize-music:promo-writer <album-name>
2. Populate promo templates manually (fill in promo/ files)
3. Skip promo review and proceed to releaseIf platform specified in arguments, review only that platform.
Otherwise, ask:
Which platforms to review?
[A] All populated platforms (campaign, twitter, facebook)
[1] Campaign strategy
[2] Twitter/X posts
[3] Facebook postsList only populated platforms as numbered options.
For each selected platform file:
## and ### headings delineate sections fences is the actual post copyFor each reviewable section, present the post with context:
## Twitter/X — Track 01: Track Title
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Track one tells the story of [concept].
Listen now: [link]
#NewMusic #HipHop
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Chars: 87/280 | Hashtags: 2 | Status: Within limits
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Actions:
[A] Approve — keep as-is
[R] Revise — give feedback for rewrite
[S] Shorten — make more concise
[P] Punch up — make more engaging/attention-grabbing
[H] Add hashtags — suggest relevant hashtags
[T] Rewrite tone — specify tone (casual, professional, hype, etc.)
[K] Skip — move to next without changesAfter any revision action (R, S, P, H, T):
Revised version:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[revised copy here]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Chars: 142/280 | Hashtags: 3 | Status: Within limits
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
[A] Approve this version
[R] Revise again — more feedback
[D] Discard — revert to originalCampaign files contain strategy tables and messaging, not code-block posts. Adjust the review approach:
After all sections for a platform are reviewed:
Updated twitter.md — 4 posts revised, 2 approved as-is, 1 skippedOnly write back if at least one section was revised. If all sections were approved as-is or skipped, skip the write step.
Between platforms, show a running summary:
## Review Progress
| Platform | Approved | Revised | Skipped | Limit Issues |
|----------|----------|---------|---------|--------------|
| Twitter | 4 | 2 | 1 | 0 |
| Facebook | — | — | — | — |
Next: Facebook (5 sections to review)
Continue? [Y/n]After all platforms reviewed:
## Promo Review Complete
| Platform | Approved | Revised | Skipped |
|----------|----------|---------|---------|
| Campaign | 3 | 1 | 0 |
| Twitter | 4 | 2 | 1 |
| Facebook | 3 | 2 | 0 |
| **Total** | **10** | **5** | **1** |
Char limit compliance: All posts within platform limits
Files updated:
- promo/campaign.md (1 revision)
- promo/twitter.md (2 revisions)
- promo/facebook.md (2 revisions)
Next steps:
1. Review any skipped posts if needed
2. Add streaming links when available (replace [Streaming Link] placeholders)
3. Ready for release: /bitwize-music:release-director <album>When revising posts, follow these principles:
[placeholder] text that needs real content| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-05 | fail→fail | 6,665 | 4,675 | -30% | 1 | 1 | 0% | 1,104 | 3,049 | +176% | 0 | 0 | — |
case-06 | fail→pass | 10,017 | 13,792 | +38% | 1 | 1 | 0% | 1,577 | 3,920 | +149% | 0 | 0 | — |
case-11 | fail→pass | 7,966 | 8,468 | +6% | 1 | 1 | 0% | 1,179 | 3,343 | +184% | 0 | 0 | — |
case-01 | fail→fail | 1,896 | 4,351 | +129% | 1 | 1 | 0% | 283 | 2,585 | +813% | 0 | 0 | — |
case-02 | fail→fail | 11,775 | 4,350 | -63% | 1 | 1 | 0% | 1,825 | 2,593 | +42% | 0 | 0 | — |
case-03 | fail→fail | 4,787 | 5,106 | +7% | 1 | 1 | 0% | 773 | 2,662 | +244% | 0 | 0 | — |
case-04 | fail→fail | 17,880 | 5,834 | -67% | 1 | 1 | 0% | 2,876 | 2,697 | -6% | 0 | 0 | — |
case-07 | fail→pass | 7,858 | 3,691 | -53% | 1 | 1 | 0% | 1,362 | 3,064 | +125% | 0 | 0 | — |
case-08 | fail→pass | 12,764 | 2,678 | -79% | 1 | 1 | 0% | 2,077 | 2,809 | +35% | 0 | 0 | — |
case-09 | pass→pass | 8,629 | 3,162 | -63% | 1 | 1 | 0% | 1,381 | 2,941 | +113% | 0 | 0 | — |
case-10 | fail→pass | 9,407 | 2,313 | -75% | 1 | 1 | 0% | 1,417 | 2,777 | +96% | 0 | 0 | — |
case-12 | pass→pass | 6,255 | 4,053 | -35% | 1 | 1 | 0% | 1,048 | 3,016 | +188% | 0 | 0 | — |
case-13 | pass→pass | 3,668 | 2,189 | -40% | 1 | 1 | 0% | 531 | 2,672 | +403% | 0 | 0 | — |
case-14 | pass→pass | 6,882 | 2,165 | -69% | 1 | 1 | 0% | 1,072 | 2,717 | +153% | 0 | 0 | — |
case-15 | pass→pass | 6,295 | 3,525 | -44% | 1 | 1 | 0% | 904 | 2,884 | +219% | 0 | 0 | — |
case-16 | pass→pass | 12,528 | 5,567 | -56% | 1 | 1 | 0% | 2,035 | 3,199 | +57% | 0 | 0 | — |
case-17 | pass→pass | 6,575 | 3,674 | -44% | 1 | 1 | 0% | 1,122 | 3,073 | +174% | 0 | 0 | — |
case-18 | fail→pass | 6,733 | 2,599 | -61% | 1 | 1 | 0% | 1,167 | 2,840 | +143% | 0 | 0 | — |
case-19 | pass→pass | 3,884 | 2,612 | -33% | 1 | 1 | 0% | 656 | 2,780 | +324% | 0 | 0 | — |
case-20 | fail→fail | 8,675 | 4,520 | -48% | 1 | 1 | 0% | 1,392 | 3,099 | +123% | 0 | 0 | — |
case-21 | pass→pass | 9,617 | 4,762 | -50% | 1 | 1 | 0% | 1,623 | 3,046 | +88% | 0 | 0 | — |
case-22 | fail→fail | 9,998 | 3,435 | -66% | 1 | 1 | 0% | 1,550 | 2,895 | +87% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +27 percentage points is the difference between those two pass rates over the 18 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.