ad-scripts
Skill6 starsChanged 4 months agoMITView on GitHub
Writes three video ad scripts in three distinct formats for one creative angle — ranked hook options, beat-by-beat spoken lines, shot notes, and CTA variants for each. Use when the angle is chosen and the scripts need to be ready to shoot or brief out.
In this repository29 other files it gives its agents.
- battlecardSkill
- brand-voiceSkill
- campaign-debriefSkill
- case-study-writerSkill
- channel-pickerSkill
- cmoSkill
- competitor-teardownSkill
- content-calendarSkill
- creative-briefSkill
- customer-deep-diveSkill
- email-sequenceSkill
- experiment-backlogSkill
- funnel-mapSkill
- hook-factorySkill
- kpi-treeSkill
- landing-pageSkill
- launch-planSkill
- lead-magnet-builderSkill
- marketing-auditSkill
- market-research-sprintSkill
- message-mapSkill
- objection-mapSkill
- offer-builderSkill
- onboardSkill
- positioning-workshopSkill
- pricing-testSkill
- repurposeSkill
- research-to-briefSkill
- weekly-reviewSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
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 34
Write for sound-off first: the on-screen text must carry the argument on its own.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
Ad Scripts — one angle, three formats
Testing three angles in one format teaches you about the angles. Testing one angle in three formats teaches you how this audience likes to be sold — and that lesson compounds across every ad that follows. This skill holds the angle constant and varies the vehicle.
Inputs
- The creative angle, the offer it sells, and the platform the ads will run on: $ARGUMENTS
- From
marketing-brief.mdif present: the language bank (voice-of-customer), top pains, the offer and its proof, objections & beliefs, and the banned-phrases list. If the angle itself is still open, run the angle work or/marketing:hook-factoryfirst — this skill takes one.
Do this
- Lock the angle. One sentence: the single belief this ad installs, and who it is for. If $ARGUMENTS carries several angles, pick the strongest, say why, and park the rest.
- Pick three formats that fit the platform and offer:
- Founder direct — one person, straight to camera, conviction and specifics.
- UGC-style — a person like the buyer, discovery register, phone-shot energy.
- Problem-demo — show the pain happening, then show the fix working. Swap in us-vs-the-old-way where the offer replaces an ingrained category habit.
- Script each in beats, spoken lines written verbatim: HOOK (first three seconds — the spoken line plus what is on screen), AGITATE (the cost of the problem, in their words), MECHANISM (why this works differently), PROOF (only what the brief can evidence), CTA.
- Write three hook options per script and rank them. The hook is most of the ad; everything after is confirmation for the people it caught.
- Add shot notes per beat — framing, setting, b-roll, on-screen text — written so someone who was never in the room can shoot it.
- Write two CTA variants per script: one direct (start, book, get) and one soft (curiosity click), matched to how cold the audience is.
- Timing pass. 30–45 seconds at natural speaking pace. Cut any spoken line the visual already says.
Output
Three complete scripts, labelled by format: ranked hooks, beat-by-beat spoken lines with shot notes, two CTA variants, target length. Close with one line per format on what a winner would reveal about how this audience buys.
Rules
- Hooks may exaggerate curiosity, never facts. No proof claim enters a script unless the brief or the user can evidence it.
- UGC-style means the register of a real person mid-discovery — not an actor reading ad copy under "authentic" lighting.
- Write for sound-off first: the on-screen text must carry the argument on its own.
- One belief per ad. A second belief is a second ad.
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.

