Slushpile

Slushpile Seven agents try to reject you before a recruiter gets the chance. What they find, you keep.

Connect on LinkedIn!

English · 简体中文 · Español · Português (BR) · Tiếng Việt · AI Bro

Install

Claude Code

claude plugin marketplace add VonTerraProject501c3/slushpile

claude plugin install slushpile@slushpile

Then, in the directory where you keep your job search:

/slushpile:onboard

Codex

codex plugin marketplace add VonTerraProject501c3/slushpile --ref main

codex plugin add slushpile@slushpile

Codex prefixes plugin skills with the plugin name:

$slushpile:onboard

Codex has no subagent dispatch, so the review pipeline runs its seven personas sequentially in one context instead of five of them in parallel. Same output, slower, and slightly more prone to one persona's reasoning bleeding into the next.

Cursor, Gemini CLI, and manual install

See INSTALL.md.

The problem

You are not being graded against the job description. You are being graded against the other seventy people who applied to the same requisition this week.

Almost every tool in this space gets that backwards. Feed a resume and a posting into a resume optimizer and it will tell you your keyword match went from 68% to 91%, which is a real number about the wrong question. If the 75th-percentile applicant in that queue matches at 94%, your 91% is a rejection, and nothing in the tool will ever tell you so.

The second thing they get wrong: they return one verdict. But the same resume and the same letter convert at maybe 2% through a cold portal submission and 30% through a referral. Those are not the same decision, and collapsing them into a single "strong match" is not a simplification. It is an error with a confident interface on top of it.

The third is the one nobody names. These tools have no memory. You paste a resume in, you get a number back, you close the tab, and the tool ends the session knowing exactly what it knew at the start. A search is forty applications over three months. Every one of them costs full price.

What this does instead

It builds a model of you, once. /slushpile:onboard interviews you and writes three files: a profile, a preferences file, and a set of stories. The profile is not a resume — it is the pool a resume gets cut from, several times longer than anything you would ever send. Every later stage reads it, and nothing asks you those questions twice.

It tries to talk you out of the role before you write anything. The search stage scores each posting against the estimated applicant pool, runs kill criteria, builds an expected-value matrix per application channel, and puts a contrarian in front of the tier list. Every other tool starts working after you have already decided to apply. The expensive mistake happens before that, and this is the only stage that can still catch it for free.