caption-craft
Skill0 starsChanged 8 days agoMITView on GitHub
Write bilingual ID/EN Instagram captions and carousel text for showing creative and technical work: relaxed, specific, never overblown, with the concrete detail doing the persuading. Use when the user wants a caption, carousel copy, or a launch post for a project, or says "buatkan caption", "caption ig", "teks carousel", "tulis caption", "write a caption". Do NOT use for ads copy, sales pages, or other people's products.
In this repository15 other files it gives its agents.
- cloudflare-workerSkill
- oss-contributorSkill
- small-briefsSkill
- councilSkill
- title-cardSkill
- repo-auditSkill
- vps-deploySkill
- interactive-figureSkill
- lab-reportSkill
- net-labSkill
- study-notesSkill
- vault-routerSkill
- html-videoSkill
- manim-explainerSkill
- proofread-idSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 47
No unnecessary emphasis about the author (how many roles they filled, that they worked alone, that it was hard).
- Distinctive: line 49
No engagement bait the author would not say out loud: no fake questions, no "tag 3 friends", no countdown drama.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Caption Craft
A caption for real work does not need to shout. The posts this skill is distilled from present builds, tools, and experiments to a mixed audience of classmates, builders, and strangers, in Indonesian and English, and the ones that work share a shape: one concrete thing, said plainly, with the numbers or the artifact visible. Hype is what a caption uses when it has no artifact.
The shape
- Hook line, one sentence, concrete. What exists now that did
not exist before, or the surprising measured result. Not a
question to the audience, not "finally released!!".
- Strong: "Tool ke-4 PerkakasID jadi: kalkulator IPK yang kemarin dibangun dua kali cuma buat membuktikan cara kedua 35% lebih hemat token."
- Weak: "Senang sekali akhirnya bisa berbagi project terbaru saya!!!"
- The middle, 2-4 short lines. What it does, how it was made, one honest limitation or lesson. Specifics over adjectives: sizes, counts, times, the stack, the failure that taught something.
- The close, one line. Where to see it (link in bio or the URL where the platform allows), or a genuine question the author actually wants answered. Never a wall of hashtags; 3-5 relevant ones at most, or none.
Bilingual rule
ID first when the audience is the author's Indonesian community, EN version below a divider for reach, and the EN is a real rewrite, not a word-for-word translation: idioms, jokes, and hooks are re-aimed, not converted. For carousel text, each slide carries one idea, in large-type-friendly short lines, and the last slide is the artifact or the link, not a summary of the summary.
Voice rules (never bend)
- Relaxed and specific. Write like the author talks about their own work: proud of the artifact, allergic to overselling it.
- No unnecessary emphasis about the author (how many roles they filled, that they worked alone, that it was hard). The work speaks.
- No engagement bait the author would not say out loud: no fake questions, no "tag 3 friends", no countdown drama.
- Claims carry their evidence or stay out. A measured number may be posted with its measurement; a feeling is phrased as a feeling.
- Emojis: a few, functional ones (a marker, a divider). A caption that needs twelve is compensating.
Variants
- Build log post: hook is the finished thing, middle is two decisions and one bug, close is the link.
- Result post (tests, benchmarks): hook is the number, middle is how it was measured and its limits, close is where the full data lives.
- Carousel: slide 1 is the hook as big text, slides 2-5 one idea each, final slide the artifact and where to get it.
Deliverable
The full caption text, ID and EN, ready to paste, plus the carousel slide text when asked. Nothing is posted by the agent; publishing is the author's tap.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

