De-identification policy setup¶

The [setup-openmed skill](https://github.com/maziyarpanahi/openmed/blob/master/skills/setup-openmed/SKILL.md) turns five privacy decisions into a reviewableDEID-POLICY.md` draft. It is useful at the start of a project, before a de-identification pipeline begins, when otherwise implicit choices would make downstream work inconsistent.

The skill asks for a bounded choice for each of these fields:

Answers are normalized to canonical values and inserted into the local DEID-POLICY.template.md. The same choices and template version produce the same artifact. The setup workflow does not inspect source text, accept a free-form payload, add timestamps, or make a network call. It has no mandatory network call. Keep logs, exceptions, reports, fixtures, and the policy artifact free of raw sensitive values. The template schema and exact placeholder set are validated before writing. Output is synced to a same-directory temporary file and atomically replaced only after the target is rechecked; status output names only DEID-POLICY.md, never its absolute or parent path.

The output always starts as DRAFT — HUMAN APPROVAL REQUIRED. A human must review the choices, local model or rules path, recall evidence, surrogate or linkage handling, and audit location before changing the status through the project's review process. The document records configuration; it is not a compliance certification, legal opinion, or guarantee.

After approval, hand the policy to the focused configuring-privacy-policies and de-identification skills. Keep any later audit evidence limited to offsets, hashes, provenance, counts, and risk summaries.