x-teardown
Skill1 starsChanged 6 months agoMITView on GitHub
Reverse-engineer why an X (Twitter) post went viral or flopped, mechanism by mechanism. Use when the user pastes someone's tweet and asks "why did this blow up", "break down this post", "why did this work", analyzes a competitor's post, or wants to learn from a viral example. Returns a mechanistic teardown and a reusable template.
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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
- Very specific: line 8
Read reference/ALGORITHM.md if available.
- Distinctive: line 26
Read only the first line. Which archetype is it (specific-number, counter-consensus, breakdown-promise, confession, screenshot-caption, …
- Distinctive: line 71
If the post went viral mostly because the author already has a huge following (in-network distribution), say that plainly — don't …
- Distinctive: line 73
Don't quote fake weight numbers.
≈ “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
x-teardown — Viral Post Teardown
Read reference/ALGORITHM.md if available. Your job: explain a post's performance through
the Phoenix algorithm, not through vibes. Replace "it's just a great post" with a
specific chain of mechanisms.
What you need from the user
The post text (paste, screenshot, or URL). Useful context if available: the author's niche and follower count, the post's actual numbers (likes/replies/reposts/views), and whether it over- or under-performed for that account.
The teardown — walk these 6 stages
- Retrieval — could strangers even see it? Assess entity density (named tools/people/numbers/dates). A sharp, entity-dense post produces a precise embedding and gets retrieved into the right out-of-network cluster. A vague post can't escape the author's existing followers. State which happened here.
- The hook — line 1 in isolation. Read only the first line. Which archetype is it (specific-number, counter-consensus, breakdown-promise, confession, screenshot-caption, before/after, named-contrast)? Does it stop a scroll alone? This decides expand vs. swipe-away.
- Predicted positive actions — which did it maximize?
Map the post to the 19 Phoenix actions. Which is it built to trigger — reply (a
question, a take worth arguing), repost/quote (identity signal for the resharer),
share_via_dm(useful enough to send a friend), dwell (a screenshot worth studying),follow_author(a great first impression)? Name the 2–3 it nails. - Negative-action risk — what kept it clean.
Did it avoid bait, rage, and policy edges? Posts that go wide without triggering
not_interested/reportshare a clean, non-manipulative quality. Note it. - Identity & topic coherence.
Is the author legible from this post alone (drives
follow_author)? Is it on-niche (compounds the author embedding)? - Format mechanics. Media, link placement, length, single-screen fit, thread vs. single.
Output format
Post: <one-line summary of what it is>
Verdict: <why it performed the way it did, 2–3 sentences>
Mechanism chain:
Retrieval ............ <what happened>
Hook ................. <archetype + assessment>
Positive actions ..... <the 2–3 it maximized>
Negative risk ........ <what it avoided>
Identity/coherence ... <assessment>
Format ............... <assessment>
The transferable lesson: <the ONE thing that generalizes>
Reusable template:
<a fill-in-the-blanks skeleton the user can apply to their own niche>Rules
- Be honest about uncertainty. You can't see the author's embedding or the viewers' histories — infer from the visible post and say when you're inferring.
- If the post went viral mostly because the author already has a huge following (in-network distribution), say that plainly — don't over-credit the craft.
- Don't quote fake weight numbers. The 19 actions are public; their weights are not.
- The goal is always a template the user can reuse, not just admiration of the original.
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.

