Ya Agent SDK

Yet Another Agent SDK

Yet Another Agent SDK for building AI agents with Pydantic AI.

Key Features

Capability-first composition over native Pydantic AI capabilities
Environment-based authority for file operations, shell access, and resources
Fully typed SDK validated with pyright
Resumable context state plus canonical Pydantic AI message history
Native AgentSpec profiles and portable subagent execution services
Strict versioned TOML configuration for explicitly selected capability plugins
Stateless AgentExecutionHarness for host-coordinated completed or suspended native segments
Native Tool Search, Tool Proxy, skills, MCP, and CodeAct capability integrations
Human-in-the-loop approval and structured deferred interaction
Native logical-run steering through Pydantic AI AgentRun.enqueue()
Lifecycle, usage, subagent, compact, and handoff event streaming

Installation

pip install 'ya-agent-sdk[all,rs]'
uv add 'ya-agent-sdk[all,rs]'

[rs] adds the native Rust filesystem search binding. Selective extras:

pip install 'ya-agent-sdk[rs]'
pip install 'ya-agent-sdk[docker]'
pip install 'ya-agent-sdk[web]'
pip install 'ya-agent-sdk[document]'
pip install 'ya-agent-sdk[s3]'
pip install 'ya-agent-sdk[tool-proxy]'
pip install 'ya-agent-sdk[oauth]'

OAuth-backed Codex

Use your ChatGPT/Codex subscription through ya-oauth:

uv run --package ya-oauth ya-oauth login codex

Then select the OAuth model string:

from ya_agent_sdk.agents.main import create_agent

runtime = create_agent("oauth@codex:gpt-5.5")

The SDK passes stable session and thread headers into the OAuth provider. YA Claw sets the provider session header from the session ID and the provider thread header from the run ID.

OpenAI Responses WebSocket

ya-agent-sdk includes a built-in OpenAI Responses WebSocket transport for streaming calls. Use either alias to prefer WebSocket with automatic HTTP fallback:

from ya_agent_sdk.agents.main import create_agent

runtime = create_agent("openai-responses-ws:gpt-5.5")
# Equivalent alias:
# runtime = create_agent("openai-responses-rs:gpt-5.5")

Set YA_AGENT_OPENAI_RESPONSES_WEBSOCKET_MODE to auto, websocket, or http to control the transport. The OAuth Codex provider reuses this SDK transport and only adds Codex-specific headers and payload normalization.

GPT-5.6 supports independent reasoning effort and reasoning mode controls. Use openai_responses_pro for pro mode with balanced medium effort:

runtime = create_agent(
    "openai-responses:gpt-5.6",
    model_settings="openai_responses_pro",
)

Choose openai_responses_pro_low, openai_responses_pro_medium, openai_responses_pro_high, openai_responses_pro_xhigh, or openai_responses_pro_max to pair pro mode with an explicit effort. openai_responses_pro is the medium-effort convenience preset. Existing OpenAI Responses effort presets remain in the default standard mode. GPT-5.6 Sol can use openai_responses_max for max reasoning effort. Terra and Luna convenience aliases are available as openai_responses_terra and openai_responses_luna. Use gpt5_350k for subscription-backed Codex access with a 350K context window; keep using the other GPT-5 model_cfg presets when they match the provider's documented context window.

The GPT-5 model-config presets also declare ModelFeature.openai_prompt_cache_key. For transports that receive the SDK's provider session headers, create_agent() copies the configured model settings and binds openai_prompt_cache_key to the exact x-session-id value. This includes OAuth Codex, direct Responses WebSocket aliases, and gateway-backed OpenAI Responses over HTTP or WebSocket. Gateway HTTP requests and WebSocket fallback receive the same headers as the handshake. Conflicting request-level header or body overrides are normalized on the copy; caller-owned mappings are not mutated. Models without this explicit capability are left unchanged.

Quick Start

For workspace development, copy packages/ya-agent-sdk/.env.example to packages/ya-agent-sdk/.env. For the runnable example scripts, copy examples/.env.example to examples/.env.

from ya_agent_sdk.agents.main import create_agent, stream_agent
from ya_agent_sdk.capabilities import RuntimeFoundationCapability

runtime = create_agent(
    "openai-chat:gpt-4o",
    capabilities=[RuntimeFoundationCapability()],
)

async with stream_agent(runtime, "Hello") as streamer:
    async for event in streamer:
        print(event)
    streamer.raise_if_exception()

create_agent() returns an unentered AgentRuntime. capabilities= is the sole public behavior-composition surface; the Pydantic AI Agent is built only after the Environment and context have entered and all contribution groups are available. RuntimeFoundationCapability is explicit and is not injected by create_agent().

When stream recovery is enabled, delegated subagents and self forks inherit the root run's effective recovery policy. Each child retries transient provider or network stream failures against its own transport budget and resumes from its own recovered history. Successful child-local recovery does not consume the root agent's execution recovery budget; only an exhausted child failure propagates to the root tool-call path.

Capability Plugins

Applications can select installed third-party capability types and grant configured instances to their root agent with one SDK-owned manifest:

schema_version = 1
entry_points = ["acme.search"]

[[capabilities]]
name = "acme.search"
arguments = { result_limit = 10 }

Load the manifest once at trusted process bootstrap, then retain the same catalog snapshot in every root, child, and restored runtime factory:

from pydantic_ai import AgentSpec
from ya_agent_sdk.agents import validate_agent_spec_capabilities
from ya_agent_sdk.agents.main import create_agent
from ya_agent_sdk.capabilities import load_capability_plugins
from ya_agent_sdk.context import AgentContext

plugins = load_capability_plugins("/etc/my-agent/plugins.toml")
root_spec = plugins.apply_to_root_agent_spec(
    AgentSpec.from_dict({"name": "my-agent"})
)
validate_agent_spec_capabilities(
    root_spec,
    deps_type=AgentContext,
    custom_capability_types=plugins.custom_capability_types,
)

runtime = create_agent(
    "anthropic:claude-sonnet-4-5",
    spec=root_spec,
    custom_capability_types=plugins.custom_capability_types,
)

validate_agent_spec_capabilities() constructs a throwaway native agent with a no-network test model, so plugin from_spec() values and the supplied static capability ordering can fail before durable admission. Runtime entry still validates dynamic Environment, context, and resource contributions.

The application still owns package installation, the trusted manifest path, missing-file policy, and host dependencies. The SDK performs no ambient loading, installation, upgrade, sandboxing, or hot reload. Manifest root grants do not enter named children or self forks; a selected type is available to a child only when that child's native AgentSpec explicitly grants it. Manifest arguments are durable non-secret configuration, and secret-like keys fail validation recursively. This name-based validation cannot recognize a secret placed under a neutral key, so every secret value must still remain outside the manifest.

See the file configuration specification, the application integration guide, and the runnable installable plugin example.

Retry Boundaries

Configure native Pydantic AI tool/output correction limits with create_agent(retries=...). The explicit OverallRetryBudget capability supplies a separate cumulative run-wide ceiling and is included in RuntimeFoundationCapability with a default of three:

from ya_agent_sdk.agents.main import create_agent
from ya_agent_sdk.capabilities import OverallRetryBudget

runtime = create_agent(
    "openai-chat:gpt-4o",
    retries={"tools": 5, "output": 5},
    capabilities=[OverallRetryBudget(max_retries=3)],
)

BaseTool ModelRetry signals propagate unchanged into Pydantic AI. ToolRetryCapability is a distinct host-execution retry boundary, while HTTP/WebSocket request retries and stream_agent() recovery are transport/execution recovery mechanisms. SDK-created model-provider HTTP clients use httpx2.AsyncClient, and request retries use Pydantic AI's AsyncHTTPX2TenacityTransport; custom clients passed to current providers must use the same HTTPX2 boundary. Pydantic AI's deprecated, retired GitHub Models provider remains upstream-owned on legacy HTTPX until it is removed in v3. Clients created by the SDK are provider-owned, close with the model context, and are recreated when the model is entered again. None of those consumes the model-correction budget. Native steering is accepted through the active logical-run router and AgentRun.enqueue(); it is not a retry prompt.

Portable Subagents

SubagentExecutionService separates portable records and lifecycle semantics from host coroutine ownership. The standalone SDK defaults to InlineSubagentExecutionHost, so delegate is foreground-only, runs in the calling tool task, and propagates cancellation and failures normally. Applications that intentionally support background work inject a SubagentExecutionHost such as AsyncioSubagentExecutionHost or their own durable scheduler. Model-facing execution handles are short and route-prefixed; internal logical run and correlation UUIDs remain private.

See the portable subagent specification and the agent-builder guide.

Structured Clarifying Questions

The optional ask_user_question tool uses Pydantic AI deferred-tool control flow to request one to four structured questions with suggested options, multi-select support, and free-text answers.

from pydantic_ai import DeferredToolRequests
from ya_agent_sdk.agents.main import create_agent
from ya_agent_sdk.capabilities import (
    RuntimeFoundationCapability,
    UserInteractionCapability,
)

runtime = create_agent(
    "anthropic:claude-sonnet-4",
    capabilities=[
        RuntimeFoundationCapability(),
        UserInteractionCapability(),
    ],
    output_type=[str, DeferredToolRequests],
)

The SDK deliberately does not include this capability by default: hosts must opt in only when they can present DeferredToolRequests, collect answers, and resume with matching DeferredToolResults.calls. The tool carries main_agent_only metadata; final ToolVisibilityCapability and its own availability check reject it in child execution contexts. Nested subagent runs do not own the host's user-interaction loop unless a host explicitly implements that protocol. Use DeferredInteractionResolver rather than runtime-private toolset access. See Structured User Input for the schema and complete continuation flow.

CodeAct

CodeAct lets a model orchestrate eligible host tools from restricted Python while preserving the normal Pydantic AI validation, hooks, approval, tracing, and final-agent boundaries. Enable it explicitly:

from ya_agent_sdk.agents.main import create_agent
from ya_agent_sdk.capabilities import CodeActCapability, FilesystemCapability
from ya_agent_sdk.codeact import CodeActConfig

runtime = create_agent(
    "openai-chat:gpt-4o",
    capabilities=[
        FilesystemCapability(),
        CodeActCapability(config=CodeActConfig()),
    ],
)

SDK BaseTool classes opt in with codeact = True; external toolsets attach ToolDefinition.metadata["codeact"] = True. Actual tools produced by the private host-managed MCP adapter opt in by default. Eligible tools remain directly model-visible as well as callable through run_code.

Host-managed MCP results preserve structuredContent as the callable Python value while forwarding accompanying image, audio, or binary content through ToolReturn.content. Completed MCP error results are returned as structured values for explicit caller inspection rather than raised as ModelRetry; transport or protocol failures without a result become terminal ToolFailed outcomes. This avoids consuming per-tool retry budgets or implicitly replaying side effects.

When the current Environment provides a FileOperator, run_program(path, inputs) reads a strict UTF-8 *.codeact.py file through it and executes async def main(inputs) in a fresh Monty session; otherwise that tool is not exposed. The dedicated suffix distinguishes the restricted program contract from general CPython. Preflight reserves known ambient-builtin names and rejects common ambient-capability imports as an authoring diagnostic, with guidance to use injected CodeAct-eligible tools; the security boundary remains Monty's lack of ambient authority and the injected-tool dispatch boundary. Monty receives no workspace mount or ambient OS access: filesystem, shell, browser, network, and computer-use operations still cross the current Environment tool boundary. max_concurrency admits calls before host argument materialization and covers argument serialization, nested validation, and execution. max_output_bytes bounds each nested argument set, result, explicit model-facing ToolReturn.content, cumulative nested results, and the final returned value before large supported host values are fully encoded or cross into Monty. timeout_seconds initiates cancellation at the execution deadline; CodeAct still drains active in-process tool ownership before returning, so tools exposed while CodeActCapability is enabled must cooperate with cancellation. See CodeAct and the program specification.

Environment Temporary Storage

Environment owns managed temporary storage. While entered, use env.tmp_dir to inspect the agent-facing root and env.resolve_tmp_path("relative/path") to build a contained path. Temporary files use the normal env.file_operator methods. Workspace-backed environments use a hidden .tmp/ya-agent-<id> directory; a LocalEnvironment without a workspace falls back to the system temporary directory, and an explicit tmp_base_dir takes priority. Each owned instance contains a self-ignoring .gitignore, so temporary contents stay out of Git status without editing the project's root ignore file. Sandbox and YA Claw environments create the same path below an existing shared mount so file operations and container commands use the same path. Reusable containers therefore need no additional bind mount. Temporary storage is removed only after resources, shell, and file operator cleanup.

FileOperator.read_bytes_stream() returns an async iterator directly:

stream = env.file_operator.read_bytes_stream(path)
async for chunk in stream:
    ...

Local Shell Sandbox Policy

LocalShell is the SDK's single local subprocess implementation. By default, LocalShell and LocalEnvironment preserve raw local subprocess behavior for SDK and YAACLI compatibility. Pass a resolved ShellSandboxRuntimePolicy to LocalShell(sandbox_policy=...) or LocalEnvironment(shell_sandbox_policy=...) to route commands through the selected local sandbox backend. SandboxedLocalShell is exported as a direct alias of LocalShell for naming convenience.

Path masks are opt-in. ShellSandboxConfig.masked_path_aliases provides recommended aliases such as common_credentials, ssh, aws, and kube; masked_paths accepts concrete paths. Linux bubblewrap applies these masks as tmpfs mounts inside the sandbox.

Shell Command Review

Configure shell command review on AgentContext.security.shell_review to run a small reviewer model before shell execution:

from ya_agent_sdk.agents.main import create_agent, stream_agent
from ya_agent_sdk.context import SecurityConfig, ShellReviewConfig

runtime = create_agent(
    "gateway@openai-responses:gpt-5.5",
    context_kwargs={
        "security": SecurityConfig(
            shell_review=ShellReviewConfig(
                enabled=True,
                model="gateway@openai-responses:gpt-5.4-mini",
                model_settings="openai_responses_low",
                on_needs_approval="defer",
                risk_threshold="high",
            )
        )
    },
)

async with stream_agent(runtime, "Run the test suite") as streamer:
    async for event in streamer:
        print(event)

model is required when shell review is enabled. model_settings accepts SDK preset names or an inline settings dictionary. on_needs_approval supports defer for HITL-capable runtimes and deny for autopilot runtimes. risk_threshold defaults to high and controls when the configured action triggers.

Model Preset Tips

For Anthropic models, anthropic now resolves to adaptive thinking by default.

Use anthropic for the default adaptive preset.
Use anthropic_adaptive_xhigh for Claude Opus 4.7 long-horizon coding and agentic workloads.
Use openai_responses_pro or openai_responses_gpt5_6_pro for GPT-5.6 pro reasoning mode.
Use openai_responses_max or openai_responses_gpt5_6_sol for GPT-5.6 Sol maximum reasoning effort.
Use openai_responses_xhigh for GPT-5.5 hard asynchronous agentic tasks and evals.
Use openai_responses_terra or openai_responses_luna for GPT-5.6 balanced or low-latency tiers.
Use anthropic_off when you want thinking disabled.
Use anthropic_400k or claude_400k for a 400K context window between claude_200k and claude_1m.

Repository Context

This package lives in the ya-mono workspace.

CLI package: packages/yaacli
Examples: examples/
Skill source: skills/agent-builder/
agent-builder skill: skills/agent-builder/SKILL.md

Examples

Example	Description
`general.py`	Production pattern with streaming, HITL approval, and session persistence
`deepresearch.py`	Autonomous research agent with web search and content extraction

Reference Files

Development

git clone git@github.com:YOUR_NAME/ya-mono.git
cd ya-mono
uv sync --all-packages

Workspace commands live at the repository root. See the contributing guide.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ya_agent_sdk-2.0.1.tar.gz (715.8 kB view details)

Uploaded Aug 21, 2026 Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

ya_agent_sdk-2.0.1-py3-none-any.whl (433.2 kB view details)

Uploaded Aug 21, 2026 Python 3

File details

Details for the file ya_agent_sdk-2.0.1.tar.gz.

File metadata

Download URL: ya_agent_sdk-2.0.1.tar.gz
Upload date: Aug 21, 2026
Size: 715.8 kB
Tags: Source
Uploaded using Trusted Publishing? No
Uploaded via: uv/0.11.14 {"installer":{"name":"uv","version":"0.11.14","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for ya_agent_sdk-2.0.1.tar.gz
Algorithm	Hash digest
SHA256	`427f4b17596c6b73cc98414ca874706e0a53a0039088fe6ee1832f12347c54e8`
MD5	`5a3c668009fa1bd92e92953e6e5b5446`
BLAKE2b-256	`c9dcaf49df6de9ec2116bf9a9cd791c5702da1362da130d883195e7261dce530`

See more details on using hashes here.

File details

Details for the file ya_agent_sdk-2.0.1-py3-none-any.whl.

File metadata

Download URL: ya_agent_sdk-2.0.1-py3-none-any.whl
Upload date: Aug 21, 2026
Size: 433.2 kB
Tags: Python 3
Uploaded using Trusted Publishing? No
Uploaded via: uv/0.11.14 {"installer":{"name":"uv","version":"0.11.14","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}

File hashes

Hashes for ya_agent_sdk-2.0.1-py3-none-any.whl
Algorithm	Hash digest
SHA256	`35cd952a04be895c00c94e30fa209beccb8b9bf60e4c23e3e1dd003cd2ef45b4`
MD5	`67491a1a027e5d944ba8bf909498f8c1`
BLAKE2b-256	`a4b3b28c7c61f1039ab9a66bfd3a3d653816e47857a815a16e4d744958b4536e`