* fix(proxy): cap managed-object poll size + expire stale rows + kill-switch flag to prevent OOM/Prisma connection loss
* fix(constants): simplify PROXY_BATCH_POLLING_ENABLED readability
* docs+test: document new polling env vars, add pagination+stale-cleanup tests
* fix: exclude stale_expired from batch poll queries; fix update_many assertions in tests
* fix: scope stale cleanup to file_purpose, fix file_object mocks, add CheckBatchCost tests
* fix: avoid duplicate cost logging in fallback path; guard integer constants against zero/negative values
* fix: cache _has_batch_processed_column; guard cleanup from aborting poll; narrow fallback except
* fix: add complete/completed to primary query not_in; fix vacuous test assertion
- Primary find_many was missing "complete" and "completed" in its not_in
filter, creating asymmetry with the fallback query. A job whose status
was set to "complete" but whose batch_processed flag update failed would
be silently re-fetched and re-processed every cycle, emitting duplicate
cost logs.
- test_fallback_completion_update_omits_batch_processed patched
_is_base64_encoded_unified_file_id to return None, causing an immediate
continue — so update() was never called and the assertion looped over an
empty list (vacuously true). Rewrote the test to mock the full
completion pipeline, verify update() is called exactly once, and assert
batch_processed is absent from the update data.
- Added symmetric test (primary path) proving batch_processed IS included
when the column exists.
Made-with: Cursor
Document the PKCE_STRICT_CACHE_MISS environment variable in config_settings.md
to fix the CI env key documentation check.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
OpenAI rejects any reasoning_effort (even 'none') with tools in
/v1/chat/completions for gpt-5.4. Update the guard to drop reasoning_effort
regardless of value. Add docs explaining the auto-drop behavior.
The existing documentation for the Responses API bridge only showed
examples with models that have `mode: responses` (like o3-deep-research),
which work automatically. This update clarifies that models with
`mode: chat` (like gpt-4o, gpt-5) require the `openai/responses/` prefix
to use built-in tools like web_search_preview.
Changes:
- Explain the `mode` property from model_prices_and_context_window.json
- List models with mode: responses vs mode: chat
- Add example showing the common error and how to fix it
- Add SDK example using the prefix with gpt-4o
- Update proxy example with both automatic and prefix-based configs
- Fix invalid trailing comma in original JSON example
Implement Anthropic Files API (upload, retrieve, list, delete, content)
using the BaseFilesConfig provider pattern. Adds multipart form-data
support to BaseLLMHTTPHandler for file uploads.
Add Claude Opus 4.6, Sonnet 4.6, Opus 4.5, Sonnet 4.5, and Haiku 4.5
to the web fetch supported models documentation. These models were
missing from the list despite supporting the web_fetch tool.
Add usage example with concrete model entry, explanation of load-time
expansion, and cross-reference to model_alias_map to clarify the
difference between the two features.
When a provider's finish_reason is mapped to a different OpenAI-compatible
value (e.g. "MALFORMED_FUNCTION_CALL" → "stop"), the original value is now
preserved in choices[].provider_specific_fields["native_finish_reason"].
This allows agent loops to distinguish between different stop conditions
without breaking the unified OpenAI-compatible finish_reason mapping.
Also returns a defensive copy from get_finish_reason_mapping() to prevent
accidental mutation of the global _FINISH_REASON_MAP.
