
Audience Persona Builder
Builds an audience profile for your account from real comment data to guide topic and script choices
- Repository
- XBuilderLAB/cheat-on-content
- Stars
- 7,443
- Licence
- MIT
- Files
- 1 file
What it does
This skill has the agent read your past post-review notes and the top comments on your published pieces, then group those comments into a profile of who actually watches and engages with your account. It saves that profile to an audience.md file, which later content-planning steps read so new ideas fit your real viewers. It is kept separate from the scoring system so it never influences how drafts are predicted to perform.
Great for
- Short-video or social creators who have several published posts with reviewed comment data
- Creators who want to check their assumed audience against what commenters actually say
- Teams using the cheat-on-content workflow who want a living audience profile refreshed over time
Skip it if
- Accounts with no published posts or review data, since the output would be guesswork (the skill itself says to refuse in that case) and benchmark-based profiles are unverified
- Anyone wanting to use the persona to score drafts, because it is designed to stay out of prediction to avoid contaminating blind scoring
- Situations needing a statistically reliable audience study, because comments only reflect the most active viewers
Try asking
- “Build persona: update my audience profile from the latest post reviews”
- “我的观众是谁? (who is my audience?) based on my comment data”
- “Refresh my audience persona after I've reviewed three more posts”
- “Show my current audience profile and what changed since the last version”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the MIT licence and an attribution note.
cheat-persona.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip cheat-persona.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip cheat-persona.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 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
- Accumulate at least three reviewed posts with comments before trusting the profile; with fewer, most traits will be labelled as hypotheses.
- Treat the output as a report of current viewers, not a target; use the gap section to decide whether to change direction yourself.
- Be aware that the skill's allowed-tools include Bash(*), so review what the agent runs in your project folder, even though the SKILL.md itself contains no scripts.
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.