Skip to content

Building AI agents, atomically

Prepare v2.13.0 (#318)

Kenny Vaneetvelde64ecf17b49e1
diff --git a/.claude/.codebase-info/.map-state.json b/.claude/.codebase-info/.map-state.json
index ba638d9..7f48e90 100644
--- a/.claude/.codebase-info/.map-state.json
+++ b/.claude/.codebase-info/.map-state.json
@@ -3,7 +3,7 @@
"tool": "codebase-mapper",
"version": "2.15.10",
"mappedAt": "2026-10-19",
- "gitCommit": "1d8ef5a32cebd9653c231062ba0bdee8e1e836eb",
+ "gitCommit": "ad832783b1e44cd7bc6a9f6e5082f17973d3b52a",
"documents": [
"architecture.md",
"coding-style.md",
@@ -17,7 +17,7 @@
"tech-landscape.md"
],
"hashes": {
- "INDEX.md": "bbe4979b8f6922896731517e748bb4775a25bad4707265f25dc821531ca545f2",
+ "INDEX.md": "0f0d07b6cf36163041d1d66ef094d566b2c3eb2f52abe7fd5151e4b261ec5266",
"architecture.md": "59d1265c89d3b42671b7c47165408625018537c016a72df421113cae3f80bb70",
"coding-style.md": "b63e8dff06d16a2d882f02624c62f490291317aec8b3032f6d79b3ee035e8567",
"communication.md": "70fbeedf239b130cfd3d7a92f27d56e96b207d85117b2f83cc9eef8ec6d6c6b3",
@@ -26,7 +26,7 @@
"entry-points.md": "3c94fd8a821fc76789794b6d3f4d89cc88a7903055755d97437f9f3ca6193cb6",
"modules.md": "cd4f81f8ce74dfc4a7bab15ba91dc569245bc342ddf494845d658b51a0a1dc51",
"onboarding.md": "3d28c671dc095a2aadb297bd7ab200c677779a99c7534e03e55a3d64c935f07e",
- "patterns.md": "dbd5ff107b7567c6424af31910ca88148836d16f31002bfed226f28b6725147f",
+ "patterns.md": "f1e2294ca48a04e9d3ce8f1e37c39e286b1b2298169004f48ceaa32cf680bb87",
"tech-landscape.md": "f7101adf177e947d5e765984326247bbbd898a46cb66bd10f49300093c391ecc"
}
}
diff --git a/.claude/.codebase-info/INDEX.md b/.claude/.codebase-info/INDEX.md
index 579078e..5c9599c 100644
--- a/.claude/.codebase-info/INDEX.md
+++ b/.claude/.codebase-info/INDEX.md
@@ -1,6 +1,6 @@
# Codebase Map — Atomic Agents
-*Last Updated: 2026-10-15*
+*Last Updated: 2026-10-19*
Atomic Agents is a lightweight, modular Python framework for building agentic AI applications as
composable, schema-driven building blocks (built on Instructor + Pydantic). This repository is a
@@ -8,7 +8,7 @@ composable, schema-driven building blocks (built on Instructor + Pydantic). This
**Stack:** Python ≥3.12 · Instructor · Pydantic v2 · LiteLLM · MCP · Textual · uv + Hatchling · Pyright · `VideoURL` multimodal content · optional OpenInference tracing (Phoenix)
**Shape:** Monorepo — `atomic-agents/` (core lib) · `atomic-assembler/` (CLI) · `atomic-forge/` (tools) · `atomic-examples/` (examples)
-**Package:** `atomic-agents` v2.12.0 on PyPI · core import package is `atomic_agents`
+**Package:** `atomic-agents` v2.13.0 · core import package is `atomic_agents`
## Documents
diff --git a/.claude/.codebase-info/patterns.md b/.claude/.codebase-info/patterns.md
index 70ea2ce..1caa64a 100644
--- a/.claude/.codebase-info/patterns.md
+++ b/.claude/.codebase-info/patterns.md
@@ -51,6 +51,7 @@ the next's input. See `atomic-examples/deep-research` and `orchestration-agent`.
clients. Core tests in `atomic-agents/tests/` mirror the package layout (`agents/`, `base/`,
`context/`, `connectors/mcp/`, `utils/`), including `tests/base/test_multimodal.py` for `VideoURL`.
Discovery (`pytest.ini`): files `test_*.py`, classes `Test*`, functions `test_*`.
+- Instructor transport regressions use real clients with `httpx.MockTransport`, including sync/async streaming and explicit strict validation.
- MCP tests use real SDK result models with mocked transport/session boundaries. Synthetic execution against production filenames is not used to increase coverage.
## Configuration
diff --git a/.release/2.13.0.md b/.release/2.13.0.md
index 53869e3..dab961c 100644
--- a/.release/2.13.0.md
+++ b/.release/2.13.0.md
@@ -1,20 +1,26 @@
# Atomic Agents 2.13.0
-Draft. Collects what has landed on main since 2.12.1.
-
## Dependencies
-- Instructor is a range again: `instructor>=1.17.0,<2`. The exact pin dated from Instructor 1.9.2 (upstream issue 567-labs/instructor#1796, an import path removed in a patch release) and was bumped by hand ever since. The framework now references `instructor.Instructor` and `instructor.AsyncInstructor` from the package root, which is the only change 1.17.0 needed; the suite passes unchanged otherwise.
-- Floors are now true floors: `pydantic>=2.11.4` (the MCP schema transformer passes `__config__` together with `__base__` to `create_model`, allowed since pydantic 2.11.4) and `mcp>=1.8.0,<2` (`mcp.client.streamable_http` first shipped in 1.8.0).
-- Two CI jobs guard this: a fresh install of the built wheel that imports every module, and a run of the suite on the lowest allowed version of every runtime dependency. Together they catch the two ways the old pin bit us: an import path removed upstream, and a floor nobody had tested.
+- Replace the exact Instructor pin with `instructor>=1.17.0,<2`. The framework uses the public `instructor.Instructor` and `instructor.AsyncInstructor` exports. (#310)
+- Set tested minimums of Pydantic 2.11.4 and MCP 1.8.0. The schema transformer needs Pydantic's support for combining `__config__` and `__base__`; the MCP connector imports a module first shipped in 1.8.0.
+- Keep `mcp<2`. Support for MCP 2.x is still pending.
+
+## Runtime fixes
-## Fixes
+- Accept Instructor's generated response schemas from its reorganized modules. Sync and async calls no longer fail the framework's docstring check on Instructor 1.17.
+- Preserve callers' `strict` setting when Instructor forwards it into streaming validation, without passing the keyword twice.
+- Exercise real Instructor request preparation and parsing through mocked HTTP transport. Tests cover sync and async calls, streaming, and explicit strict settings in both streaming paths.
-- `agent.run()` and `run_async()` work on Instructor 1.15.3 and later. Instructor moved its generated response-model wrapper to `instructor.v2.core.function_calls`, and `BaseIOSchema` only exempted the older module names from its docstring rule, so every non-streaming call raised before the request went out. The check now accepts any class Instructor generates.
-- `run_stream()` with `model_api_parameters={"strict": ...}` no longer fails on Instructor 1.17 ("got multiple values for keyword argument 'strict'").
-- Both are covered by a new test module that drives a real Instructor client through an `httpx.MockTransport`, so Instructor's own request and parse path runs in the suite instead of a mocked client.
+## Installation checks
+
+A new CI job runs the suite on the lowest compatible direct runtime dependencies. The clean-wheel import check introduced in 2.12.1 remains in place.
## Compatibility
-- `max_retries` counts differently on Instructor 1.17: 1.14.5 stopped after `max_retries` attempts in total, 1.17.0 stops after the first attempt plus `max_retries` retries. `max_retries=3` passed through `model_api_parameters` now makes up to four calls instead of three. Lower it by one to keep the old budget.
-Python 3.12 or newer. Applications that pin `instructor==1.14.5` keep working; applications that move to 1.17.0 need no code change on their side. openai 3.x resolves and passes the suite with Instructor 1.17.0.
+- Python 3.12 or newer.
+- This release requires Instructor 1.17 or newer within the 1.x line. Applications that need Instructor 1.14.5 can stay on Atomic Agents 2.12.1.
+- Instructor 1.17 changes integer `max_retries` counting. A value of 3 now permits the first attempt plus three retries, rather than three attempts total. Lower it by one to preserve a previously configured attempt budget. Atomic Agents forwards this setting without changing it.
+- No live model-provider calls were used for release verification. The locked and minimum-dependency suites each passed with three credential-gated tests skipped.
+
+Full changelog: https://github.com/Eigenwise/atomic-agents/compare/v2.12.1...v2.13.0
diff --git a/pyproject.toml b/pyproject.toml
index 734894d..9c606cf 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
[project]
name = "atomic-agents"
-version = "2.12.1"
+version = "2.13.0"
description = "A versatile framework for creating and managing intelligent agents."
readme = "README.md"
license = "MIT"
diff --git a/uv.lock b/uv.lock
index d3132eb..b062de4 100644
--- a/uv.lock
+++ b/uv.lock
@@ -267,7 +267,7 @@ wheels = [
[[package]]
name = "atomic-agents"
-version = "2.12.1"
+version = "2.13.0"
source = { editable = "." }
dependencies = [
{ name = "gitpython" },