encode-policy-v2

name	encode-policy-v2
description	Use when the user invokes $encode-policy-v2 or asks Codex to implement a new PolicyEngine-US state benefit program from official rules. Covers research, source collection, requirement extraction, scoped implementation, tests, validation, and draft PR preparation.

encode-policy-v2

This is the Codex equivalent of the Claude /encode-policy-v2 command.

Treat the text after $encode-policy-v2 as the command arguments:

$encode-policy-v2 STATE PROGRAM [--skip-review] [--research-only] [--600dpi]

Before starting, read workflow.md and subagents.md. Follow them as the source of truth for phases, checkpoints, handoff files, validation gates, and Codex subagent delegation.

Operational rules:

• Use /tmp/{PREFIX}-... handoff files to keep large research, diffs, PDFs, and reports out of the main context.
• Ask the user before making scope decisions that affect modeled requirements.
• Use PolicyEngine skills explicitly when implementing: $policyengine-us, $policyengine-parameter-patterns, $policyengine-variable-patterns, $policyengine-testing-patterns, $policyengine-code-style, $policyengine-code-organization, and $policyengine-vectorization.
• This workflow is designed for Codex subagents. When subagent use is available and authorized, delegate independent phases to worker or explorer agents using subagents.md; otherwise execute the same steps directly while preserving handoff-file boundaries.
• Every PDF reference href must include #page=XX, using the PDF file page number.