Skip to main content
Core targets the complete OpenAI Agents API as pinned below (public API rule). This ledger records how much of each resource Core implements and which contract holds its details, then every known difference from the OpenAI service and every open gap. API namespaces and credentials says who calls which API; the Agents API guide shows how to use it.

​
Pinned baseline

Contract tests hold Core to the pin: the router and openapi.yaml serve exactly the pinned routes (services/core/internal/api/routing_test.go, v1/upstream_contract_test.go), every query parameter and field is official, and Core-only fields sit only inside x_agents_core on Agents and Sessions. Swagger 2.0 cannot express string-or-array unions, so openapi.yaml leaves Session input and function-result output unconstrained; the pinned types and Core’s validation define them. Operations and fields newer than the pin wait for a protocol upgrade. Evidence for a status comes from the pinned official SDK and raw HTTP against the running service, as CONTRIBUTING requires.

​
Coverage by resource

Implemented means every operation serves the pinned shapes; limits that remain are listed under known gaps. Partial names what is missing. Which operation each Harness supports on each placement is in the Harness capabilities. Core wire behavior holds the rules that apply across resources: requests, errors and lists. Core’s own fields sit inside x_agents_core (Core extensions). The Core administration API (/core/v1) and the machine API (/api/v1) are not part of the Agents API.

​
Differences from OpenAI

Each item is Core’s deliberate or native behavior where the official service behaves otherwise. The linked rule states the exact behavior. Requests and errors (Core wire behavior)
  • Core sends no OpenAI-Organization or OpenAI-Project response headers.
  • HEAD on the event stream, on content downloads and on the Environment files list returns 405.
  • A JSON array body is rejected; the official service reads [] as {}.
  • U+0000 in a stored string returns 400; the official service stores it.
  • Not-found messages never name the resource; Core quotes a full metadata key where the official message abbreviates it.
  • UUID identifiers also resolve in other spellings, such as uppercase or braces.
  • The Files routes keep a local unsupported_parameter code for a repeated query key.
Lists (lists)
  • A deleted Agent or Session used as a cursor returns 404; the official service still pages from it.
  • A Turn cursor from another Session, a Credential cursor equal to its Vault ID, and a Skills cursor that is not a Skill ID return 404.
  • Vault and Credential lists clamp a negative limit, as the pinned SDK describes; the official service returns 400.
Agents and Sessions (Agents, Sessions)
  • A repeated Session creation with the same Idempotency-Key returns the original Session; the official service creates a new one.
  • Omitted programmatic tool calling keeps the harness’s native behavior; the official default is on.
  • An omitted reasoning effort stays null instead of taking the model’s default.
  • A Session’s agent.tools omits tool_search declarations.
  • Deleting a Session right after an events 202 returns 409, because Core admits the Turn in the same transaction; the official service returned 200.
Input, events and history (Sessions, events and history, message content)
  • Input to a none Session is admitted synchronously; Core does not emulate the official asynchronous admission window.
  • A function result that resumes a waiting Turn emits turn.in_progress, and cancelling a Turn that waits on a function result emits an interim agent.session.in_progress.
  • Attaching to the stream mid-Turn sends no catch-up Item snapshots.
  • The Items list includes in-progress and incomplete output Items, and keeps a failed function result’s submitted output.
  • Error messages omit the call and executor IDs that official messages include.
  • An empty text part beside other text is accepted and stored.
  • Empty input returns 400 invalid_request with a generic message and a null param; the official response is invalid_request_error with param input.
  • Session usage is available as soon as every root Turn has settled; official reads lag by seconds.
Files, Skills, Environment files and Artifacts (Files and Skills, Environment files and Artifacts)
  • File uploads hold up to 512 MiB; the official limit is 512 MB. The Files list returns up to 10,000 Files by default, and a purpose filter other than user_data returns an empty page.
  • Skill version numbers are never reused, and uploads and deletions of one Skill run one at a time.
  • Environment files work on self_hosted Environments, which the official service refuses.
  • Creating an Environment file over an existing regular file returns the “must not traverse symlinks or overwrite existing files” message. A parent symbolic link that stays inside the workspace is followed, and a parent that escapes the workspace or is a regular file gets a generic 400; the official service rejects symbolic-link parents.
  • Artifact IDs are UUIDs.
Vaults and Credentials (Vaults and Credentials)
  • Vault and Credential status is stored privately and defaults to active; with no archive operation, lists without a filter include both statuses.
  • An unknown or foreign Vault in vault_ids returns 404 “Resource not found.”; the official message names the ID.
  • An explicit null for OAuth access_token, refresh or token_endpoint_auth in an update keeps the stored value.
  • A static token must be an RFC 6750 b64token to run; other stored tokens fail at dispatch.
  • Vault metadata has a 64 KiB bound and no pair or length limits; names are 1–256 bytes after trimming.

​
Known gaps

Configuration and tools
  • Explicit reasoning effort or summary, service tiers other than auto, enabled web_search and enabled programmatic tool calling are saved but rejected at Session admission.
  • Harness support for tools, structured output, deferred discovery, subagents and MCP differs by Harness and placement; see the Harness capabilities. MiniMax Code has no public functions, no service-origin MCP and no image input.
  • Model-derived reasoning defaults are not resolved.
Execution and history
  • The stream does not emit reasoning-summary events, Environment pending or ready events, or every pinned interim tool-output variant.
  • Native Item variants beyond those listed under Turns and Items are not projected, and Items cannot be modified.
  • A function result that cancellation prevents from being applied never appears as an Item.
  • Pinned Codex can lose command output emitted before its stream subscription.
  • Claude Code and MiniMax Code report no public usage.
  • Core gives no crash-safe or exactly-once guarantee for native side effects; claimed work fails on restart without replay.
  • Images must be inline PNG or JPEG data URIs; remote URLs, file_id and detail are rejected.
Environments and Templates
  • Runtimes do not enforce disabled or restricted networks, so Sessions that need them are rejected (restricted network policy).
  • packages.system is rejected; system packages must be preinstalled.
Files and Environment files
  • Files accept only purpose=user_data; other purposes, expires_after and the Uploads API are not supported.
  • An Environment file write whose outcome is uncertain is never retried or recovered automatically; it blocks further writes and messages to the Session.
Vaults and Credentials
  • There is no archive lifecycle, storage-key rotation or re-encryption.
  • OAuth refresh happens only at dispatch: there is no refresh on a provider 401, no mid-Turn replacement and no withdrawal of a token already sent to a Runtime.
  • Input sent to a Session whose selected Credential was deleted is admitted and then fails at dispatch.
  • Credential creation takes no Idempotency-Key.
Sessions
  • Session deletion does not purge the stored history physically.
  • Stream lifetimes for self_hosted, hosted and no-input creation, and the creation-stream retry, are Core’s own choices.

​
Unverified against the official service

  • The order of errors when one request has several faults, and error, default and payload-limit parity in general.
  • Whether Subagent child Turns and pending Environment file writes block Session deletion.
  • The official order between the pending-input error and an unknown result target.
  • Failure reasons for npm, initial-file and Skill installation have no official sample.
  • Codex behavior for an empty text part beside text, and for images in failed function results or as remote references; Claude behavior for a whitespace-only or empty text block inside a mixed message.
  • Official HEAD behavior on the routes where Core returns 405.
  • Environment Template hostname forms beyond exact hosts, and disabled combined with domains.
  • Files purpose filtering and pagination while Files change; official Skill upload limits and error timing.
  • Environment file list defaults (limit 20, the workspace root as the default path, no recursion, page-token invalidation), the 50 MiB file_id copy limit and the check order on create.
  • Artifact capture of hard links, special files and a linked outputs directory, republication after changed bytes, and content headers and ranges.
  • Vault and Credential error and retry semantics, pagination under concurrent writes, visibility after deletion, exact-URL matching against the official normalization, OAuth refresh timing and errors, and restricted-key scopes.