pick-creative-mechanic
Skill0 starsChanged 27 days agoMITView on GitHub
Choose, name or rebuild the creative mechanic of an ad (the move that makes the viewer reach the point themselves, between angle and format) from a fixed vocabulary matched to awareness stage, with built examples from the AdSwipeFile ad library. Use when a concept has an angle and a format but no structural idea, when explaining why an ad lands, or when a brief needs more than "say this in this format".
In this repository14 other files it gives its agents.
- audit-ad-creativeSkill
- competitor-adsSkill
- creative-strategySkill
- find-winning-adsSkill
- generate-ad-conceptsSkill
- get-startedSkill
- map-competitorsSkill
- map-personasSkill
- record-creative-learningSkill
- review-saved-adsSkill
- teach-ad-breakdownSkill
- track-competitor-adsSkill
- weekly-competitor-reviewSkill
- write-hooks-from-adsSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 44
Not every ad has one. Plain offer ads and direct demos often run on the claim alone: say none rather than forcing a fit.
- Very specific: line 50
Start from the concept's angle, stage and trigger moment (generate-ad-concepts).
- Distinctive: line 51
Ask how this persona should reach the truth: see it, overhear it, be told by someone else, have a belief turned, compare it.
- Distinctive: line 53
Ask what the user already knows about their own ads: which moves they have run, and what their results showed.
- Distinctive: line 76
Never state an ad's results: spend, ROAS, conversions, or that one ad beat another.
≈ “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
Pick a creative mechanic
Needs the AdSwipeFile connector. This skill reads the AdSwipeFile ad library of Meta ads through its MCP server. If no AdSwipeFile tools (
search_library,list_ads) are available, say so before anything else and give the step for this host. Claude Code:claude mcp add --transport http adswipefile https://adswipefile.com/api/mcp. claude.ai or ChatGPT: addhttps://adswipefile.com/api/mcpas a custom connector. Other clients: https://adswipefile.com/docs/mcp?utm_source=agent-skill&utm_medium=referral&utm_campaign=pick-creative-mechanic. Connecting opens a sign-in to an AdSwipeFile account. Steps that need only the user's own material can go on; never invent ads, brands or links in place of the library.
The angle is what the ad argues; the format is how it is produced; the mechanic is the move that
gets the viewer to the angle's truth. A mechanic travels across formats: the same move works as a
talking head, a static or a carousel. A concept with a format and no mechanic is a production spec;
with a mechanic and no format it is an idea nobody can shoot. Words are used as the
creative-strategy skill's glossary defines them.
The vocabulary
Most mechanics work by letting the viewer draw the conclusion, which skips the "this is an ad" reflex. Use 0 to 2 per ad, main first. They are the names creative strategists already use, so a brief reads without a key.
| Slug | The move | Stages |
implied-answer | A question or setup the visual answers without anyone saying it | unaware, problem aware |
social-witness | Someone else notices the change (a compliment, a double take); the product shows up as the reason | problem aware, product aware |
overheard-conversation | Framed as a private exchange (texts, DMs, a call) the viewer overhears, where one person recommends it | unaware, problem aware |
reframe | Puts the problem, the price or the product in a new frame: the real cause, cents a day, a different job | problem aware, solution aware |
common-enemy | Names an outside enemy (an industry, an ingredient, a habit) and sides with the viewer against it | solution aware, product aware |
trojan-horse | Dressed as something that is not an ad (useful content, a story, a news item); the product arrives late as the resolution | unaware |
contrast-without-comment | Two states side by side with no verdict spoken; the gap makes the case | problem aware, solution aware |
this-and-a | The product placed among everyday things the viewer already loves or owns, so it inherits the feeling | unaware, problem aware |
demonstration | The product proven under the test the viewer doubts (the spill, the run, a result measured on camera) | solution aware, product aware |
mirroring | Describes the viewer's own habits or moment so precisely they recognise themselves | unaware, problem aware |
objection-handling | A real doubt named in so many words (too expensive? does it work?) and answered | product aware, most aware |
cost-of-inaction | Shows what going on without it costs: money, time, health, a missed moment | problem aware, solution aware |
Not every ad has one. Plain offer ads and direct demos often run on the claim alone: say none rather than forcing a fit. A move outside this list gets a one-line definition and a new slug, never the nearest name.
Picking one
- Start from the concept's angle, stage and trigger moment (
generate-ad-concepts). - Ask how this persona should reach the truth: see it, overhear it, be told by someone else, have a belief turned, compare it. Keep the mechanics whose stages include the concept's.
- Ask what the user already knows about their own ads: which moves they have run, and what their results showed. A move their results back outranks a fresh one; one that fell flat needs a reason to try again. Their results come from them (pasted, a report, another connector).
- Find two or three built examples on AdSwipeFile:
list_adswithmechanicset to the slug, the concept'sawarenessstage andniche, andsort: "longest_running". Too few: dropniche, thenawareness; for a move outside the list,search_librarywith its plain words ("text message", "before after", "comment reply"). Thenget_ads(up to 25 per call) and read eachbreakdown. They show the construction, never a result. - Pair at most two: one primary that shapes the concept, one secondary that adds a layer
(
social-witnessunderimplied-answer;reframeinsidetrojan-horse).
Output, in the reply: the mechanic slug (primary, optional secondary), one line on how this concept
executes it, the stage fit, and the reference ads by url. It goes into the concept's brief.
Naming the mechanic of an existing ad
For an audit or a breakdown: name what the viewer has to do to get the point, say why it slips past
ad skepticism, then check whether the ad completes it (does the visual answer the question? does the
rival read as the rival?). For a library ad, read it with get_ad and start from its breakdown
and why_it_works. An incomplete mechanic is a finding for audit-ad-creative.
Rules
- Never state an ad's results: spend, ROAS, conversions, or that one ad beat another. A long run is a reason to study an ad: write "kept running for 94 days".
- Link every cited ad by its
url. - The library tags mechanics with these slugs, so
list_adswithmechanicfinds them, but an ad's result doesn't show its tag: a mechanic you name for one library ad is your reading of itsbreakdown, so say so.
Keep the examples
Offer, once, to put the examples on the concept's board (create_board or one from list_boards,
then add_ad_to_board) with a set_ad_note naming the mechanic each shows. Only on a yes.
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.

