3.2: principal layer + tiered reporting core SDK support

[bokelley]

Python twin of the TypeScript issue (adcp-client); protocol PRs adcp#6966, adcp#7020, adcp#6953.

1. Regenerate Pydantic models from the 3.2 RC bundle (surfaces renamed since the prototypes: sync_principal/get_principal, adcp.principal, principal_id, principal.changed, reporting tier flags).
2. Principal layer client + server primitives: sync/get helpers, version-fence bootstrap, destination setup polling, declarations negotiation readback; server-side principal record, proof state machine hooks, suspension/revocation, principal.changed emission.
3. Reporting core reconciler reworked from prototype feat(reporting): prototype reliable ledger reconciliation #1095 to the tiered model: Core path free of destination/manifest/digest/receipt code; optional tiers feature-gated.
4. Testing fixtures mirroring the protocol's reporting-core fixture and universal/principal storyboard.

🤖 Generated with Claude Code

[bokelley]

Triage

Classification: Feature request (3.2 epic)
Bucket(s): types-codegen, client, handlers — no exact label matches found (gap flagged in run summary)
Status: ready-for-human

What the experts said:

• ad-tech-protocol-expert: Epic-shaped; gated on adcp#6966/7020/6953 merging and schema bundle publishing; tiered reporting approach is protocol-sound, but proof state machine / suspension-revocation hooks need validation against adcp#6953's state diagram before item 2 is scoped.
• code-reviewer: Epic-shaped; renamed symbols (sync_principal, adcp.principal, principal_id, principal.changed) are breaking public API changes requiring aliases.py and the lazy __init__ facade to be updated; PR feat(reporting): prototype reliable ledger reconciliation #1095 only blocks item 3, not items 1/2/4.

My take: All four items are gated on the 3.2 RC protocol PRs (adcp#6966, adcp#7020, adcp#6953) merging and the schema bundle publishing to the canonical registry — CLAUDE.md policy prohibits regenerating from anything but the published source. This is genuinely epic-sized and should be converted to an epic with a task list, sequenced as: codegen regen → principal client+server → reporting reconciler (after #1095 progresses) → fixtures, with separate sequential PRs to avoid codegen traversal-order churn cascading across simultaneous branches.

Recommend @bokelley:

• Convert to epic / milestone tracking with task list
• Gate item 1 on adcp#6966, adcp#7020, adcp#6953 all merging and the RC schema bundle publishing
• Review adcp#6953's principal state diagram before scoping item 2's proof state machine / suspension-revocation hooks
• Sequence item 3 after feat(reporting): prototype reliable ledger reconciliation #1095 resolves or is superseded

---

Triaged by Claude Code. Session: https://claude.ai/code/${CLAUDE_CODE_REMOTE_SESSION_ID}

---

Generated by Claude Code