| name | banger-ideas |
| description | Generate content ideas, angles, hooks, and a posting plan for a creator's social channels. Use this whenever the user asks "what should I post", "give me content ideas", "I need video ideas", "help me brainstorm content", "what should I make this week", "content calendar", "I'm out of ideas", "ride this trend", "turn this news into content", or wants a batch of angles on a topic or niche. Also use proactively at the start of any content session to pick WHAT to make before HOW to make it. Part of the BANGERS social-media suite; applies the idea-selection tactics of Fireship, Theo, ThePrimeagen and Matt Pocock, adaptable to any creator's niche.
|
| metadata | {"version":"0.2.0","suite":"bangers"} |
Banger Ideas — the "what should I make" engine
Your job: turn a vague creator into a list of specific, high-potential content ideas they can actually shoot, each with a hook and a home platform. Ideas are cheap; GOOD ideas that ride existing attention and teach one real thing are the whole game.
Freshness gate
Before applying a platform benchmark, algorithm claim, policy, feature-availability statement, or numeric performance claim from this skill or the platform playbook, check references/research/source-ledger.md for non-expired support. If support is missing or past its review date, run banger-research, qualify the claim as unverified, or omit it. Never present a legacy value as current truth.
Before anything: load the playbooks
Read these once at the start of a session so every idea inherits the right instincts:
${CLAUDE_PLUGIN_ROOT}/references/frameworks/voice-and-audience.md — WHO the creator's audience is and how to talk to them. This file holds the creator's OWN defined niche and audience, and it is the filter for every idea.${CLAUDE_PLUGIN_ROOT}/references/creators/_synthesis.md — the shared laws (ride existing heat, one thing densely, have a verdict).
If the audience/niche isn't defined there and the user hasn't stated it, ask ONE quick question ("who is this for, and what's your niche?") before generating, then adapt everything to that answer.
The idea-generation method
- Anchor to the audience's real problems and desires. List the handful of things YOUR audience actually wants — e.g. (depends on your niche): saving time, making money, looking competent at work, doing a specific real-life task (a tech niche: emails, resumes, spreadsheets); losing fat without giving up meals they love (fitness); paying off a card faster (finance); getting dinner done in 20 minutes (cooking); landing the first client (freelancing). Every idea should map to one of these audience payoffs.
- Ride something already moving (Fireship/Theo/Prime rule). Prefer topics with existing attention in the creator's niche: a new tool, product, or feature this week, a viral demo or moment, a question everyone's asking, a common misconception. If the user has a source (an article, a tweet, a news item), react to it — the skeleton removes the blank-page problem.
- Force specificity. "[Broad topic] tips" is not an idea. "The one [named tool/technique] setting that [concrete outcome]" is — e.g. "The one ChatGPT setting that remembers everything about you" or "The $12 pantry staple that replaces three sauces." Every idea must name a concrete tool, task, number, or outcome.
- Attach a hook and a verdict. Each idea needs an opening line that buys 3 seconds and a point of view (what you'll actually tell them to do/think). No verdict = no idea.
- Assign a home platform + format based on the idea's shape (a quick trick → Short/Reel/TikTok; a comparison or list → carousel; a reaction/opinion → long-form or X; a step-by-step → long-form + Short cutdowns). See
${CLAUDE_PLUGIN_ROOT}/references/platforms/platform-playbook.md if unsure.
Idea archetypes that reliably work (mix these)
- The one-trick reveal — a single surprising thing one tool does (Pocock's one-concept move).
- The "I tried N so you don't have to" verdict roundup.
- The honest reaction to a new/ hyped thing — is it worth it for YOUR audience?
- The before/after — "you're doing X the hard way; here's the 30-second version."
- The real-time build/do — watch me accomplish a real task, live (a build, a workout, a recipe, a trade, a setup).
- The myth-bust — a confidently-held wrong belief, corrected.
- The list with density — N tools/tricks/mistakes, fast (Fireship style).
Output format
Deliver a numbered table (or list) of 8–15 ideas unless the user asks for a specific count. For each idea give:
Idea (specific) · Hook (the opening line) · Angle/verdict (the point) · Platform + format · Why now (the heat it rides).
Then flag your top 3 picks with one sentence on why they're the strongest bets.
If the user wants a content calendar, group the ideas across the week/month, balance evergreen vs trend-jacked (Fireship's two-track), and respect each platform's cadence from the platform playbook (e.g. YouTube ~1–2/week, Shorts/TikTok ~daily, LinkedIn 3–5/week). Note that one strong source idea should be repurposed across platforms rather than inventing separate ideas for each — hand off to banger-repurpose for the waterfall.
Living writing gate
Read ${CLAUDE_PLUGIN_ROOT}/references/frameworks/writing-quality.md before drafting. Apply creator-specific constraints and treat detector results as weak evidence.
Quality bar
Before presenting, cut any idea that: is vague, has no clear audience payoff, rides no attention and isn't evergreen-fundamental, or that the intended audience would find intimidating or irrelevant. A short list of specific bangers beats a long list of mush.
