Policies for sources, checks, and response requirements | Andy's AI Playbook
Policies for choosing sources, checking claims, reviewing code or system design, and setting response requirements.
Choose a policy by task
Choose the policy that matches the control you need: sources, claim checks, engineering review, response requirements, research workflows, or the default technical baseline.
Choose sources
Policies for deciding where facts may come from and what kinds of sources are allowed.
Use when: Use this group when the first question is where the answer is allowed to get its facts.
Prompt files (1) Hide prompt files
Check claims and writing
Policies for verifying factual claims, technical terminology, and publication-ready technical writing.
Use when: Use this group when the output is already drafted and you need to check whether it is accurate, precise, or ready to publish.
Prompt files (1) Hide prompt files
Prompt files (2) Hide prompt files
Prompt files (2) Hide prompt files
Review code and system design
Policies for architecture review, boundary review, and implementation review against official sources.
Use when: Use this group when the task is an engineering review of code, configuration, architecture, or dependency structure.
Prompt files (2) Hide prompt files
Prompt files (2) Hide prompt files
Set response requirements
Policies for controlling how the response should look, sound, and report confidence.
Use when: Use this group when the main requirement is about the output contract rather than source choice or review.
Prompt files (1) Hide prompt files
Prompt files (1) Hide prompt files
Research and literature review
Policies for structured literature review, evidence synthesis, source comparison, and gap mapping.
Use when: Use this group when the task is a scholarly review rather than a normal answer or engineering check.
Recommended baseline
Cross-cutting default rules for objective, technical, non-simulative output.
Use when: Use this as the default baseline when no narrower policy overrides it.