Agent Reflection guides
How to write rules your agents can use, how to test and group them, how to run the critique loop in your agent, and how to read what comes back.
Writing rules
How do I turn a style guide into a rule my agents can use?
Pick one concern, write it the way you would brief a careful reviewer, and save it as a draft rule. Your agent can draft the Markdown and upload it for you.
How do I keep my brand guidelines in one place instead of in every agent?
Split the guidelines by concern, save one rule for each, group them in a rule set, and update the set when the guidelines change.
How do I test a rule before my agents rely on it?
Use the test status, run one output that should pass and one that should fail, then set the rule to active.
How do I check an output against several rules in one call?
Pin rules to versions in a rule set. One call checks the output against all of them, for the price of one.
Running critiques in agents
How do I run the critique loop in Claude Code?
Connect Claude Code to Vindex, ask it to read your rules, and let it check and revise a draft until every rule passes or it reaches the attempt cap.
How do I put the critique loop in my agent's standing instructions?
Add one short instruction to the file your agent reads on every task, so it knows to check its output against your rules on each task.
Reading results and reports
How do I read a critique result?
What pass or fail, the quoted passage, the attempt count, and your credits left mean, and what to do with each refusal.
How do I find the rule my agents fail most?
Ask your agent for the failure rate by rule and by rule version, and read what the figures mean.
Not connected yet?
Every guide assumes your agent can call Vindex. The first-run page takes a few minutes, and the docs define every term. Would rather start from a finished rule? Browse the rules marketplace. Looking for story scoring in your assistant? See the AI client guides.