You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
epic: Python SDK parity with JS @adcp/sdk@6.7 framework primitives #452
Building the v3 reference seller (PR #447) confirmed adopters hand-roll patterns that JS now ships as helpers. The JS 6.7 migration guide at docs/migration-6.6-to-6.7.md (in adcp-client-js) documents 15 recipes.
Goal
Ship the high-leverage 11 helpers as Python SDK primitives so adopters (Prebid salesagent, GAM-shaped publishers, etc.) get the same DX as JS.
Classification: Feature request (epic — multi-batch, cross-cutting) Bucket(s): handlers, client, validation (labels not present in repo — gap noted below) Status: ready-for-human Milestone: (omitted — no target version specified)
Assessment: Epic is well-structured. Batch 1 (4 parallel items) feeds Batch 2's dependency on AccountStore.upsert/list/sync_governance(ctx) from item #4; Batch 3 is independent DX polish. PR #447 (v3 reference seller motivating this) merged this morning. The Batch 1 item #4 (Account v3 wire fields + AccountStore) is the critical-path gate — Batch 2 cannot start until it lands.
My take: Keeping this as a tracking epic is correct. When child issues are filed, triage will process each one individually — Batch 1 items are non-breaking additions and Execute-eligible; Batch 2/3 items will be evaluated as each child lands. No auto-PR from this issue.
Label gap:handlers, client, and validation bucket labels do not exist in this repo. Adding them would enable label-based filtering across epic children.
Closing the tracking epic: the two remaining children called out in the last status update (#453 and #462) are now closed as completed. Any new parity gaps should be filed as focused follow-up issues rather than keeping this stale epic open.
Activity
bokelley commentedon May 3, 2026
Triage
Classification: Feature request (epic — multi-batch, cross-cutting)
Bucket(s): handlers, client, validation (labels not present in repo — gap noted below)
Status: ready-for-human
Milestone: (omitted — no target version specified)
Assessment: Epic is well-structured. Batch 1 (4 parallel items) feeds Batch 2's dependency on
AccountStore.upsert/list/sync_governance(ctx)from item #4; Batch 3 is independent DX polish. PR #447 (v3 reference seller motivating this) merged this morning. The Batch 1 item #4 (Account v3 wire fields + AccountStore) is the critical-path gate — Batch 2 cannot start until it lands.My take: Keeping this as a tracking epic is correct. When child issues are filed, triage will process each one individually — Batch 1 items are non-breaking additions and Execute-eligible; Batch 2/3 items will be evaluated as each child lands. No auto-PR from this issue.
Label gap:
handlers,client, andvalidationbucket labels do not exist in this repo. Adding them would enable label-based filtering across epic children.Triaged by Claude Code. Session: https://claude.ai/code/${CLAUDE_CODE_REMOTE_SESSION_ID}
Generated by Claude Code
feat(server): createRosterAccountStore — Shape C explicit AccountStor…
feat(decisioning): createOAuthPassthroughResolver — Shape B account r…
feat(decisioning): createOAuthPassthroughResolver — Shape B account r…
feat(decisioning): create_tenant_store — opinionated multi-tenant Acc…
feat(decisioning): create_tenant_store — opinionated multi-tenant Acc…
feat(client): looks_like_v3_capabilities — drop the v2-downgrade foot…
feat(server): createMediaBuyStore — opt-in targeting_overlay echo for…
feat(validation): oneOf near-miss validator hints + issues[].hint on …
feat(validation): oneOf near-miss validator hints + issues[].hint on …
feat(decisioning): wire credential strip into dispatch + webhook + re…
feat(validation): oneOf near-miss validator hints + issues[].hint on …
feat(decisioning): wire credential strip into dispatch + webhook + re…
feat(decisioning): credential-leak strip on every echo path + ctx_met…
bokelley commentedon May 14, 2026
Status check — 9 of 11 children shipped
Bringing the checkboxes up-to-date based on a cross-check against merged PRs:
Batch 1 (parallel, no deps)
create_upstream_http_client+create_translation_map— still open (clarify status; 3 code-reviewer blockers from triage)compose_method+ security composersref_account_id— closed (shipped via feat(decisioning): state-machine helpers + typed exception classes + ref_account_id #465)AccountStore.upsert/list/sync_governance(ctx)— closedBatch 2 (depends on #456) — complete
create_tenant_storecreate_oauth_passthrough_resolver— shipped via feat(decisioning): createOAuthPassthroughResolver — Shape B account resolver factory #472 (issue closed today)create_roster_account_store— shipped via feat(server): createRosterAccountStore — Shape C explicit AccountStore for publisher-curated rosters #471 (issue closed today)Batch 3 (DX polish)
oneOfnear-miss validator hintslooks_like_v3_capabilitiesclient fallback heuristiccreate_media_buy_store— still open (clarify status; 3 correctness blockers from triage)Remaining work: #453 and #462 — both need design decisions before they can draft. Leaving this epic open until both close.
Validation: v3 reference seller has been refactored to use most of these helpers in subsequent PRs.
bokelley commentedon May 23, 2026
Closing the tracking epic: the two remaining children called out in the last status update (#453 and #462) are now closed as completed. Any new parity gaps should be filed as focused follow-up issues rather than keeping this stale epic open.