fix(protocol): support AdCP 3.1.15 schemas

[bokelley]

Summary

• bump the bundled protocol pin from AdCP 3.1.14 to 3.1.15
• verify the published bundle checksum and Sigstore identity
• regenerate cached schemas and Pydantic exports
• record that 3.1.15 introduces no SDK model field-shape changes

Upstream 3.1.15 backports compliance grading for rate-limit exhaustion; the SDK-facing schema shape is unchanged. This will release as SDK 7.0.2 because 7.0.1 is already published and immutable.

Validation

• make regenerate-schemas
• make validate-generated
• pytest -q tests/test_schemas_version_pin.py tests/test_code_generation.py tests/test_adcp_3_1_14_schema_backport.py — 24 passed
• pre-commit: black, ruff, mypy, adopter fixtures, bandit, JSON and secret checks all passed
• upstream bundle checksum and Sigstore release-workflow identity verified

[aao-ipr-bot]

⚠️ Argus review could not complete

The automated review encountered an issue (possibly reached max turns, timed out, or failed to post the final gh pr review). A human reviewer should take this PR.

View workflow run

_{This is an automated message from the Argus AI review workflow.}

[aao-secretariat]

Ladon verdict: Approve

Approve — routine protocol pin bump AdCP 3.1.14 → 3.1.15.

The reviewer found no actionable issues. Every bundled schema under schemas/cache/** changed only its generatedAt timestamp; registry files bumped the version string; the two generated_poc/** files changed only the codegen timestamp comment and $schema example strings inside docstrings. No enum members, required[] entries, discriminators, or field types moved, and no anonymous-variant renumbering occurred — so aliases.py coverage remains intact. The fix(protocol): prefix correctly maps to a patch since the public surface is unchanged.

Decision-table walk:

• Row 1 (critical/high): no findings — skip.
• Row 2 (gated_paths): gated_paths is false — skip.
• Row 3 (high-risk deleted): all high-risk entries are (modified), none (deleted) — skip.
• Row 4/5 (medium on schema/infra or modified high-risk): no findings at all — skip.
• Rows 6–8: no prior escalation, no no-auto-approve team match, zero medium findings — skip.
• Row 9: approve.

high_risk is true only because the diff touches schemas/** (all modifications, no deletions), and the reviewer confirmed those modifications are timestamp-only churn preserving every contract. No escalation evidence exists.

[aao-secretariat]

Ladon verdict: Approve

Approve — clean schema/version re-pin, no blocking or medium findings.

This is a metadata-only delta since the prior approve (head cb2c792). The reviewer confirms tests/conformance/signing/vector_manifest.json was re-pinned from AdCP 3.1.14 → 3.1.15 (spec_version + source), with vector content hashes unchanged — correctly reflecting that 3.1.15 leaves the request-signing surface untouched and only backports rate-limit-exhaustion grading. The pin now agrees with src/adcp/ADCP_VERSION. The remaining diff is regenerated schema-cache churn under schemas/**.

Decision table walk:

• Rows 1: no critical/high findings.
• Row 2: gated_paths is false — does not apply (REVIEW_REQUIRED status is moot).
• Row 3: high_risk true but no (deleted) reasons — all (modified).
• Rows 4–5: no medium findings, so the high-risk-modified escalation does not fire.
• Row 6: prior decision was approve (not escalate) — sticky rule N/A.
• Row 7: no no-auto-approve team match.
• Row 8: fewer than three medium findings (zero).
• Row 9: approve.

High-risk flag is driven entirely by (modified) schema-cache regeneration output; with no reviewer concerns at any tier, these modifications are presumed safe.