Skip to content
Get packs

Cheat Bump: Rubric Upgrade

Proposes and runs guarded upgrades to a content-scoring rubric or its bucket boundaries

Source Download skill
Repository
XBuilderLAB/cheat-on-content
Stars
7,443
Licence
MIT
Files
1 file

What it does

This skill teaches an agent to manage a personal content-scoring formula (a rubric that rates posts). When you ask to change the formula's weights or dimensions, the agent expands your request into a full equation, re-scores past posts through a separate blind sub-agent, checks whether the new ranking matches real view counts, and asks a cross-model reviewer to audit the result before anything is saved. A lighter mode only re-derives the score bucket boundaries without touching the formula.

Great for

  • Creators who keep a prediction log and want to recalibrate their content scoring against real performance
  • Users of the cheat-on-content toolkit who already have rubric_notes.md, predictions files and state files set up

Skip it if

  • People who want a general content idea or writing tool: it does nothing without the surrounding cheat-on-content project files
  • Small sample sizes or noisy data: the skill itself warns that conclusions from few samples are weak, and a rubric change based on them can mislead

Try asking

  • “bump rubric: change the engagement weight from 1.5 to 2.0 and drop the NA dimension”
  • “--bucket-only: my account has grown, recalibrate the score buckets”
  • “我想加一个维度 to track memetic shareability”
Step-by-step guide: how to use Cheat Bump: Rubric UpgradeInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    cheat-bump.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip cheat-bump.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip cheat-bump.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give exact numbers in your proposal (for example ER 1.5 to 2.0); the skill refuses to guess vague requests like 'raise ER a bit'.
  • Expect the skill to abort if the sub-agent tool is unavailable; use --bucket-only for offline work instead.
  • Treat the rank-consistency result as internal only: the skill itself says the final decision must wait for the cross-model audit.

What's inside

  • SKILL.md

Redistributed unmodified from XBuilderLAB/cheat-on-content under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.