fix(api): pop null json_schema description/strict omit-real before proxy by cursor[bot] · Pull Request #646 · ContextualWisdomLab/contextual-orchestrator

cursor · 2026-08-16T16:26:24Z

Summary

SDK optional defaults serialize omitted response_format.json_schema.description / strict as JSON null or blank.
Accepting those keys without popping them is not omit-equivalent: proxy_completion forwards the body and several providers reject strict: null.
Pop null/blank description and null strict in place so passthrough matches omit.
Fail closed on unknown inner keys and non-string description (invalid_response_format).
Docs: caller next-action notes plus APA 7th cites for OpenAI Chat Completions, Structured Outputs, and IETF JSON Schema 2020-12.
Unique tip on the fix(api): pop null tool.function optional fields omit-real before proxy #638 omit-real substrate. Do not merge this stack onto main. Land the unique commit after Full unit + Semgrep.

Test plan

python3 tests/test_json_schema_inner_fields_omit_http_honesty.py (ok)
python3 tests/test_tool_function_null_fields_pop_http_honesty.py (ok)
python3 tests/test_chat_response_format_http_honesty.py (ok)
python3 tests/test_tool_strict_null_noop_http_honesty.py (ok)
python3 tests/test_responses_response_format_http_honesty.py (ok)
paper / API / plugin-driven / product-planning contracts (ok)
CI Full unit + Semgrep

Reviewer next action

Review the unique tip (d59f164) only. Do not APPROVE the 143-file honesty stack as a main merge. Independent non-author APPROVE is still required for the unique commit.

…closed otherwise Chat history: message-level audio and legacy function_call are null/empty omit no-ops; non-empty fail closed with named errors (including tools passthrough). Tip substrate from #577 assistant refusal/annotations honesty. Local full unit: 940 passed.

…ed otherwise OpenAI fine-tune style message weight is not applied on this gateway. Accept null/0/1 as honest no-ops; reject other types and values with invalid_message_weight. Tip substrate from #578. Local full unit: 943 passed.

…ion role Reject unsupported message keys with named unknown_message_fields (not silent strip or tools-passthrough smuggle). Reject legacy function role with invalid_message_role migration to tool. Tip substrate from #579. Local full unit: 947 passed.

OpenAI partial-assistant prefix flag is not applied on this gateway. null/false are honest no-ops; true and non-booleans fail closed with invalid_message_prefix. Tip substrate from #580. Local full unit: 950 passed.

…therwise Named invalid_max_tool_calls on /v1/chat/completions instead of opaque unknown_fields. Aligns with Responses max_tool_calls honesty; gateway has no multi-step tool loop.

…losed otherwise Legacy /v1/completions treated max_tool_calls as unknown_fields. Accept the key for named invalid_max_tool_calls (null/empty/whitespace omit-equivalent), matching chat/Responses honesty so SDKs get a clear migration path.

SDK clients often send include_usage/include_obfuscation as JSON null. Drop null flag values before validation so null (and null+false mixes) match omit / all-false no-ops on chat, Completions, and Responses. True flags remain fail-closed with invalid_stream_options.

…or Responses parallel true SDK optional defaults often send function.strict and json_schema.strict as null — treat as omit rather than type errors. Align Responses parallel_tool_calls=true with chat by requiring a non-empty tools array.

SDK optional defaults often send description and parameters as JSON null. Treat null as omit rather than type errors; non-null non-string/object values remain fail-closed with invalid_tools.

OpenAI-style tool descriptions are at most 1024 characters. Over-long descriptions fail closed with named invalid_tools so SDKs never believe a truncated description was accepted.

SDK optional participant name blanks ("" / whitespace) are omit-equivalent like JSON null. Non-string, over-long, and invalid charset names remain fail-closed with invalid_message_name.

SDK optional defaults serialize omitted tool.function description/parameters/strict as JSON null. Accepting those keys without popping them is not omit-equivalent: proxy_completion forwards the body and several providers reject null parameters. Pop in place so passthrough matches omit; keep non-null wrong types on invalid_tools. Also pop response_format.json_schema.strict null. Tip substrate from #614. Local full unit: 989 passed.

…n keys SDK optional defaults serialize omitted response_format.json_schema description/strict as JSON null or blank. Leave those keys and the gateway forwards them; several providers reject strict: null. Pop in place so passthrough matches omit. Unknown inner keys and non-string descriptions stay invalid_response_format. HTTP echo tests cover chat and Responses. Docs cite OpenAI structured outputs and IETF JSON Schema 2020-12 in APA 7th. Co-authored-by: Seongho Bae <seonghobae@users.noreply.github.com>

SDK optional defaults serialize omitted function descriptions as empty or whitespace-only strings. Leaving those keys is not omit-equivalent: proxy_completion forwards them and several providers reject a blank description. Pop in place so passthrough matches omit; keep non-empty strings and fail-closed non-string values. Co-authored-by: Seongho Bae <seonghobae@users.noreply.github.com>

Co-authored-by: Seongho Bae <seonghobae@users.noreply.github.com>

cursor

Stale comment

Unique tip review (d59f164 + 1cb13ae + b09aac3 on #638 569f760)

Verdict: SOUND for the omit-real slice. Do not merge this 151-file honesty stack onto main.

What the unique tip does

Pops JSON-null / blank response_format.json_schema.description and JSON-null strict in place so proxy_completion forwards omit, not null.

Fail-closes unknown inner keys and non-string description with named invalid_response_format.

Pops empty / whitespace-only tool.function.description the same way (SDK blank default).

HTTP echo.response_format / echo.tools contracts cover chat and Responses.

Local evidence

python3 tests/test_json_schema_inner_fields_omit_http_honesty.py
python3 tests/test_tool_description_blank_omit_http_honesty.py
python3 tests/test_tool_function_null_fields_pop_http_honesty.py
python3 tests/test_chat_response_format_http_honesty.py
all printed ok.

Residual (next unique tip, not a merge of this stack)

json_schema.name is only “non-empty string.” OpenAI Structured Outputs require [a-zA-Z0-9_-]{1,64} — the same charset already enforced on tool.function.name. Invalid names still reach the provider and the buyer sees an opaque upstream error instead of invalid_response_format. Do not fold that onto this PR; land it as the next unique commit after Full unit + Semgrep on this tip.

Reviewer next action

Review the three unique commits only. Independent non-author APPROVE is still required. Do not APPROVE the 143-file substrate as a main merge. CI Full unit + Semgrep are queued; they are not a reason to stop the next honesty slice.

_{Sent by Cursor Automation: Fix Issues}

cursor · 2026-08-16T16:32:58Z

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.

cursor

Stale comment

Unique tip review (569f760..b09aac3)

Unique commits d59f164 + 1cb13ae + b09aac3 are SOUND. Omit is real: _validate_chat_response_format pops JSON-null / blank response_format.json_schema.description and JSON-null strict on the same dict proxy_completion forwards. Unknown inner keys and non-string descriptions stay invalid_response_format. Blank tool.function.description is popped after the #638 null helper.

Local scripts printed ok:

tests/test_json_schema_inner_fields_omit_http_honesty.py

tests/test_tool_description_blank_omit_http_honesty.py

tests/test_tool_function_null_fields_pop_http_honesty.py

tests/test_chat_response_format_http_honesty.py

tests/test_responses_response_format_http_honesty.py

tests/test_tool_strict_null_noop_http_honesty.py

Reviewer next action

Do not APPROVE or merge this 151-file honesty stack onto main. Review the unique tip only. Independent non-author APPROVE is still required for the unique commit.

Residual (later unique tip, not this PR)

Official Responses structured output is text.format. This gateway still returns 400 invalid_text for any non-empty text — fail-closed, not a null leak. The Responses HTTP test locks the chat-style response_format shim only. Next honesty slice: json_schema.name must match [a-zA-Z0-9_-]{1,64} (tool names already do). Do not fold #639/#640 mode/tool_choice here. Envelope work stays on #628/#648. Instructions omit stays on #649.

CI Full unit + Semgrep were queued at review time — not treated as a blocker.

_{Sent by Cursor Automation: Fix Issues}

cursor

Unique tip review (`569f760`..`b09aac3`)

Unique commits d59f164 + 1cb13ae + b09aac3 are SOUND. Omit is real: _validate_chat_response_format pops JSON-null / blank response_format.json_schema.description and JSON-null strict on the same dict proxy_completion forwards. Unknown inner keys and non-string descriptions stay invalid_response_format. Blank tool.function.description is popped after the #638 null helper.

Local unique + related honesty scripts printed ok on this run's substrate. Code-reviewer subagent confirmed the same evidence (server.py 2442 / 2484 / 2523-2539; mock _mock_raw echoes the nested dicts).

Reviewer next action

Do not APPROVE or merge this 151-file honesty stack onto main. Review the unique tip only. Independent non-author APPROVE is still required for the unique commit.

Residual (later unique tips, not this PR)

json_schema.name [a-zA-Z0-9_-]{1,64} is already on #654. Do not fold it back here.
Official Responses structured output is text.format (flat type/name/schema/strict). This gateway still returns 400 invalid_text for any non-empty text (server.py 1183-1197). Chat-style response_format on /v1/responses is not that coverage. Next honesty slice: accept official text.format with the same omit-real / fail-closed rules.
Do not fold #639/#640 mode/tool_choice here. Envelope work stays on #628/#648. Instructions omit stays on #649.

CI Full unit + Semgrep were still queued at review time — not treated as a blocker. Name-charset thread on this PR is fully addressed by #654.

_{Sent by Cursor Automation: Fix Issues}

seonghobae · 2026-08-16T18:21:19Z

Superseded by tip substrate ≥ #691 (cumulative OpenAI/gateway honesty band + auto-merge tip). Closing to free product-gate runners (Full unit + Semgrep).

seonghobae and others added 15 commits August 16, 2026 21:18

fix(api): treat message prefix null/false as omit; fail-closed on true

82038d3

OpenAI partial-assistant prefix flag is not applied on this gateway. null/false are honest no-ops; true and non-booleans fail closed with invalid_message_prefix. Tip substrate from #580. Local full unit: 950 passed.

fix(api): treat chat max_tool_calls null/empty as omit; fail-closed o…

1a196b0

…therwise Named invalid_max_tool_calls on /v1/chat/completions instead of opaque unknown_fields. Aligns with Responses max_tool_calls honesty; gateway has no multi-step tool loop.

fix(api): treat tool.function description/parameters null as omit

5dace0f

SDK optional defaults often send description and parameters as JSON null. Treat null as omit rather than type errors; non-null non-string/object values remain fail-closed with invalid_tools.

fix(api): cap tool.function.description at 1024 characters fail-closed

55d7a45

OpenAI-style tool descriptions are at most 1024 characters. Over-long descriptions fail closed with named invalid_tools so SDKs never believe a truncated description was accepted.

fix(api): treat chat message name empty/whitespace as omit

4f84700

SDK optional participant name blanks ("" / whitespace) are omit-equivalent like JSON null. Non-string, over-long, and invalid charset names remain fail-closed with invalid_message_name.

docs(api): note blank tool.function.description omit on the pop helper

b09aac3

Co-authored-by: Seongho Bae <seonghobae@users.noreply.github.com>

cursor Bot commented Aug 16, 2026

View reviewed changes

Comment thread contextual_orchestrator/server.py

cursor Bot mentioned this pull request Aug 16, 2026

fix(api): fail-closed json_schema.name charset and length #654

Closed

6 tasks

seonghobae marked this pull request as ready for review August 16, 2026 16:32

seonghobae self-requested a review as a code owner August 16, 2026 16:32

seonghobae enabled auto-merge (squash) August 16, 2026 16:32

cursor Bot commented Aug 16, 2026

View reviewed changes

cursor Bot mentioned this pull request Aug 16, 2026

fix(api): accept official Responses text.format omit-real #657

Closed

10 tasks

cursor Bot commented Aug 16, 2026

View reviewed changes

cursor Bot mentioned this pull request Aug 16, 2026

fix(routing): fail-closed ablation θ and persist streamed snapshots #662

Closed

seonghobae mentioned this pull request Aug 16, 2026

fix(api): accept official Responses text.format type=text #666

Closed

9 tasks

cursor Bot mentioned this pull request Aug 16, 2026

fix(api): accept official Responses text.format structured types #681

Closed

12 tasks

seonghobae closed this Aug 16, 2026

auto-merge was automatically disabled August 16, 2026 18:21
Pull request was closed

Navigation Menu

Uh oh!

fix(api): pop null json_schema description/strict omit-real before proxy - #646

fix(api): pop null json_schema description/strict omit-real before proxy#646
cursor[bot] wants to merge 15 commits into
mainfrom
cursor/bc-7c35c4a5-ec8a-4cb2-a81e-4555bb06f521-d286

cursor Bot commented Aug 16, 2026

Uh oh!

cursor Bot left a comment •

edited

Loading

Uh oh!

Uh oh!

cursor Bot commented Aug 16, 2026

Uh oh!

cursor Bot left a comment •

edited

Loading

Uh oh!

cursor Bot left a comment

Uh oh!

seonghobae commented Aug 16, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

cursor Bot commented Aug 16, 2026

Summary

Test plan

Reviewer next action

Uh oh!

cursor Bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Unique tip review (d59f164 + 1cb13ae + b09aac3 on #638 569f760)

What the unique tip does

Local evidence

Residual (next unique tip, not a merge of this stack)

Reviewer next action

Uh oh!

Uh oh!

cursor Bot commented Aug 16, 2026

Uh oh!

cursor Bot left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Unique tip review (569f760..b09aac3)

Reviewer next action

Residual (later unique tip, not this PR)

Uh oh!

cursor Bot left a comment

Choose a reason for hiding this comment

Unique tip review (569f760..b09aac3)

Reviewer next action

Residual (later unique tips, not this PR)

Uh oh!

seonghobae commented Aug 16, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

cursor Bot left a comment •

edited

Loading

Unique tip review (`d59f164` + `1cb13ae` + `b09aac3` on #638 `569f760`)

cursor Bot left a comment •

edited

Loading

Unique tip review (`569f760`..`b09aac3`)

Unique tip review (`569f760`..`b09aac3`)