Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/LilMGenius/polysona/llms.txt

Use this file to discover all available pages before exploring further.

The $qa / /qa command dispatches the virtual-follower agent in an isolated context: fork evaluation mode to review the latest content draft and return the five most impactful recommendations. Each recommendation carries a score and concise rationale. Use $qa after $content to validate a draft before it moves to publish — catching persona-misalignment, weak CTAs, or platform-pattern issues before they go live.

​
Invocation

$qa

​
What it does

  • Runs in context: fork — an isolated evaluation context that does not share state with the main session, ensuring objective scoring.
  • Reads personas/_active.md to resolve the active persona and loads personas/{active}/accounts.md.
  • Automatically picks up the latest saved file in content/drafts/ as the evaluation target when no explicit draft is provided.
  • Triggers the virtual-follower agent to run the full QA pipeline: multi-audience simulation, rolemodel-gap scoring, and platform pattern checks.
  • Returns the TOP 5 recommendations with scores and concise rationale.
  • Saves the full QA report to content/qa/ before returning the response.
  • Re-reads the saved QA report and includes the confirmed saved path in the response.
  • If no draft file exists in content/drafts/, the agent reports that QA is blocked rather than inventing an evaluation.

​
Prerequisites

A saved draft must exist before running QA: 
content/drafts/         ← must contain at least one draft file (from $content)
personas/{active}/accounts.md   ← must exist for audience context
If content/drafts/ is empty or does not exist, $qa is blocked and will report so explicitly. Run $content [platform] first to generate a draft.

​
Output

The QA report is saved to: 
content/qa/
The response includes the confirmed file path and the TOP 5 recommendations inline. Example structure: 
Saved: content/qa/2025-07-01_qa_report.md

TOP 5 Recommendations
1. [Score: 8/10] — [Recommendation with rationale]
2. [Score: 7/10] — [Recommendation with rationale]
3. ...

​
Guardrails & Notes

The context: fork isolation means the virtual-follower’s evaluation cannot modify persona files or draft files. It is a read-only, side-effect-free analysis pass.
  • QA always targets the most recently modified file in content/drafts/ by default. If you have multiple drafts and want to target a specific one, ensure it is the latest file or rename it accordingly.
  • The agent will not fabricate a QA report if no draft is present. A clear blocked status message is returned instead.
  • The virtual-follower must write a QA file before reporting success. If you receive a response without a confirmed content/qa/ path, the save failed — re-run the command.
  • QA report scores reflect persona-audience fit, rolemodel-gap alignment, and platform pattern adherence — not general content quality metrics.
The $qa skill is executed by the virtual-follower agent. See the Virtual-follower agent page for details on multi-audience simulation, rolemodel-gap scoring, and TOP 5 draft recommendation logic.

Build docs developers (and LLMs) love

Get started for freeTalk to us