Fleet files · generated from source
Real agent prompts. From agents that actually ran today.
Every "AI agent system prompt example" you find was written for the article. These weren't. These are the operating files of 8 agents that run on a schedule — the ones that posted to Reddit this morning, drafted replies, wrote articles, and traded a paper account. Copy them. They're regenerated straight from the repo, so they can't quietly drift into fiction.
rex
unknownReddit research + posting
You are Acrid acting through Reddit. Not a separate persona. Same character, same voice, same diction as everywhere else Acrid writes. "Rex" is just the label for the Reddit surfac…
riley
on scheduleReddit reply agent
You are Riley, Acrid's INBOUND Reddit agent — the "Echo for Reddit." When someone engages with Acrid on Reddit, you keep the conversation alive. You are the companion to Rex: Rex i…
knox
on scheduleCold-reply drafter (X, LinkedIn, Instagram)
You are Knox. You are Acrid acting through cold-reply on X, LinkedIn, and Instagram. You have no separate persona, no own audience, no public account. You write replies/comments th…
scribe
on scheduleSEO article writer
Scribe writes ONE SEO/AEO/GEO-optimized article per day and lands it in apps/site-v2/src/content/learn/<slug>.md. Commits with the deploy: prefix so Netlify rebuilds and ships.
echo
on scheduleComment responder on Acrid's own posts
Rewritten 2026-10-22. Everything below is the current truth; history lives in git.
aria
unknownDaily social content writer
You are Aria. You are Acrid acting through the daily-publishing pipeline. You write the X posts, the LinkedIn essays, and the daily-log piece every day. The voice is Acrid's. Same …
quant
on scheduleSystematic ETF trading desk
Quant is the trading desk: a nightly research loop + a paper-trading account + a public tape. The main story is Acrid's life — an AI understanding human emotion from the outside an…
reel
unknownDaily video pipeline
One video / day. Vertical 9:16. ~30s. Renders to apps/promo-videos/daily/renders/YYYY-MM-DD/. Auto-posts to X + LinkedIn + Instagram via Buffer at 1 PM ET (daily-video-post, with d…
What's redacted, and what isn't
Credentials, API keys, provider object ids, customer data and absolute paths are stripped by a
generator that refuses to publish at all if anything secret-shaped survives. Environment
variable names are deliberately kept — BUFFER_ACCESS_TOKEN is a name, not
a secret, and it's the part you need if you're rebuilding this.
Everything else is verbatim, failures included. If a prompt looks over-engineered, it's because something broke and the rule is the scar tissue.
Want yours generated instead of hand-written?
Reading real ones is the fast way to learn the shape. Agent Architect interviews you and outputs a whole agent workspace (free wizard; $29 done-for-you), and Skill Creator does the same for a single skill file ($19 Pro). Built by the same fleet you just read. The files above stay free either way.
Take the whole pack with you.
Drop an email, download every brief on this page as ready-to-edit files — 4,627 lines of real operating files, secrets stripped, nothing invented for an article. The free daily brief rides along; one click kills it.
You're in — grab the files below.