No. It picks targets, writes and lints drafts, and tells you how to measure. You or your own tooling post. No API keys needed.

Will this work on my account?

The numbers come from one small account over two days (104 replies), so treat them as direction. Step 7 shows how to log your own replies and update the rules for your audience.

What does reply_check.py check?

It blocks over-length drafts, links in reach replies, advice-shaped replies, hedges and empty praise, a first line that depends on the parent post, and LinkedIn comments that open with a reaction. It warns on AI-sounding phrasing and forecasts views at about 0.6% of the parent's views.

Is this engagement bait or spam?

No. It is about writing one good reply under the right thread, with every fact checked, no links in reach replies, and no punching at private people.

Does it cover LinkedIn?

Yes, a comment variant: comment within the first 60 to 90 minutes and open with a claim the reader can use, not a reaction to the post.

What is in the download?

SKILL.md (target scoring, line-1 test, openers, never-list, fact check, measurement loop, LinkedIn variant, examples) and scripts/reply_check.py (standard library Python).

Viral Reply Writer for X and LinkedIn

Example session with this skill installed

Here's a thread I want to reply to: a Nintendo post about physical game cartridges, 98 minutes old, ~4K likes, ~300K views. Top replies are all 'finally' jokes. Write my reply.

Read your context and instructions
Compiled the viral reply writer

Target check

Factor	Value	Verdict
Parent views	~300K	good (250K+)
Age	98 min	late; still worth one reply
Reply room	'finally' jokes only	open angle: nostalgia
Forecast	~0.6% of 300K ≈ 1,800 views	worth it

Drafts

every kid who blew on a cartridge in 1995 now owns a shelf, and nintendo just told them it's safe
the only thing dying faster than the cartridge was the idea that you own your games / nintendo just slowed one of them down
three generations of gamers blew into a cartridge to fix it / none of them ever found out if it worked

Picked #1: crowd opener, line 1 needs zero context, nobody in the replies has the nostalgia angle.

Lint

$ python3 scripts/reply_check.py --parent-views 300000 "every kid who blew on a cartridge in 1995 now owns a shelf, and nintendo just told them it's safe"
LINE 1: every kid who blew on a cartridge in 1995 now owns a shelf, and nintendo just told them it's safe
  + crowd opener ("every kid who...")
forecast: ~1800 views (about 0.6% of the parent's 300000; wide spread, treat as order of magnitude)
VERDICT: OK TO POST (after the fact check)

Fact check

Claim implied: Nintendo announced continued physical releases. Verified in the parent post itself. No numbers used.

Next steps

Post it now (age matters more than polish).
No link in this reply. Log it; check views after 6 hours and add the row to your table.

Connects securely to your tools. The creator never sees your data.

About this skill

What it does

On a small account your own posts reach almost nobody. Replies under big live threads borrow the thread's audience. This skill is the measured version: which threads, which first lines, which shapes, and what never to post, plus a linter that blocks the known failure shapes before you hit send.

The data behind it (real, one account)

An agent team ran a small X account (under ~120 followers) for two days and analyzed every reply and post it made (104 replies, 28 posts, plus LinkedIn). Medians:

Replies vs the account's own posts: 668 vs 52 views (~13x)
First line that works with zero context vs one that needs the news: 1,316 vs 162 (~8x)
Parent thread 250K-1M views vs under 50K: 2,129 vs 141 (~15x)
A reply gets about 0.6% of the parent's views (so you can forecast before writing)
Openers: reversal 3,222 and crowd 2,516 vs plain quips 641 vs tips/advice 65
Best single reply: 87,917 views
LinkedIn: early comments on big posts got 4,059 and 5,673 impressions vs the account's own posts at 126-294

Small sample, one account; the skill tells you how to measure your own and update the rules.

What you get

Target scoring: size, age (under 30-60 min), reply room (likes-to-replies), topic, safety, and X search queries that find live threads
The line-1 stranger test and the two openers that won (crowd: "every kid who..."; reversal: then vs now)
The never-list (tips, links, hedges, name-dependent punchlines, invented facts, punching at private people)
A fact-check step and a 6-hour measurement loop
LinkedIn comment variant (timing + open with a usable claim)
reply_check.py: blocks over-length, links, advice shape, hedges, a line 1 that depends on the parent, reaction-first LinkedIn comments; warns on AI-sounding phrasing; forecasts views from parent size

No API keys, no auto-posting, no engagement bait.

How to use

Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.

Reviews

No reviews yet

Be one of the first to try it. Every listed skill passes our trust checks below.

Security scanned

Passed our 8-point scan before listing

Fresh listing

Recently published to Agensi

30-day refund

Not a fit? Get your money back

Viral Reply Writer for X and LinkedIn