chore(main): release 0.2.0

[github-actions]

🤖 I have created a release beep boop

0.2.0 (2026-06-25)

Features

• agent_api_keys: one-call webhook-trigger setup endpoint (#329) (9cc71fb)
• agentex-ui: grow chat input vertically for multi-line prompts (#332) (8c060f1)
• retention: strict bool env parsing + stale-RUNNING cleanup override (#306) (5ad81e8)
• task-messages: add optional is_error to ToolResponseContent (#331) (5020e96)
• update task configs in agentex db on turn (#309) (13eea7a)

Bug Fixes

• authz: grant legacy agent register principal (#325) (cae5f94)
• deps: cap fastapi <0.137.0 to stop OPTIONS preflight 500s (#334) (acbe9e3)
• deps: clear golden-image Trivy CRITICAL/HIGH (litellm, starlette, pyjwt, python-multipart) (#320) (bfa6652)
• streams: cut SSE error-log volume and add Redis pool headroom (#340) (352eaaa)
• streams: snapshot SSE cursor before "connected" to stop first-token drops (#330) (7c2ccfd)
• tasks: preserve task_metadata in combined update and forward merge_params by name (#336) (8c25c64)

Documentation

• guides: tracing, framework-agent, and local-sandbox guides (#339) (b43e8c3)
• streaming: document the unified harness surface (#337) (31fd697)

---

This PR was generated with Release Please. See documentation.

Greptile Summary

This PR prepares the 0.2.0 release. The main changes are:

• Bumps package and API version metadata to 0.2.0.
• Adds the generated 0.2.0 changelog.
• Refreshes the checked-in OpenAPI spec.

Confidence Score: 4/5

The release metadata changes are straightforward, but the lockfile and telemetry defaults need attention before merging.

The changed files are limited and the identified issues are concrete release/versioning problems with clear remediation paths.

pyproject.toml, agentex/src/_version.py

T-Rex Logs

What T-Rex did

• Regenerated the frozen lockfile and ran the reproduction script with uv 0.7.3, which showed pyproject packages at 0.2.0 while uv.lock still recorded 0.1.0 for agentex packages, executed uv lock --check which failed with exit status 2, and exercised the Dockerfile release commands uv export --frozen --no-dev --no-emit-project --package agentex-backend and uv sync --frozen --no-dev --package agentex-backend, with export and sync completing but the stale-lock validation reproducing the failure.
• Reproduced the telemetry version defaults by running a focused Python harness with SERVICE_VERSION unset, which imported src._version.__version__ as 0.2.0 while init_otel_metrics created telemetry Resource service.version as '0.1.0' and otel_metrics.get_meter was called without a version, returning version '0.1.0'.
• Observed a version delta: the base state had 0.1.0 across root pyproject, agentex pyproject, _version.py, and OpenAPI info.version, with a successful exit, while the head state had 0.2.0 across the same files, also with a successful exit.

_{Ran code and verified through T-Rex}

Prompt To Fix All With AI

Fix the following 2 code review issues. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 2
pyproject.toml:3
**Regenerate frozen lockfile**

This release bumps the workspace package to `0.2.0`, but `uv.lock` still records both `agentex` and `agentex-backend` as `0.1.0`. The Docker image copies that lockfile and runs frozen `uv export` / `uv sync` for `agentex-backend`, so the release build can fail the frozen-lock check or install/export package metadata for the previous version instead of this release.

### Issue 2 of 2
agentex/src/_version.py:3
**Update telemetry version defaults**

The API version now comes from `__version__` and is `0.2.0`, but `src/utils/otel_metrics.py` still falls back to `SERVICE_VERSION` `0.1.0` and `get_meter(..., version="0.1.0")`. Internal callers use `get_meter` without passing a version, so deployments that do not set `SERVICE_VERSION` emit 0.2.0 service metrics under the old 0.1.0 version, which can make rollout dashboards and version comparisons incorrect.

_{Reviews (11): Last reviewed commit: "chore(main): release 0.2.0" | Re-trigger Greptile}

Greptile also left 2 inline comments on this PR.

[greptile-apps]

Regenerate the lockfile

The package versions were bumped here and in agentex/pyproject.toml, but uv.lock still records both workspace packages as 0.1.0 (agentex and agentex-backend). The Docker build copies that lockfile and runs uv sync --frozen --no-dev --package agentex-backend and uv export --frozen, so the release image path can fail the frozen-lock freshness check or install/export package metadata from the previous version. Regenerate and commit uv.lock with the 0.1.1 package versions.

Artifacts

Repro: lock metadata check script

• Contains supporting evidence from the run (text/x-python; charset=utf-8).

Repro: lock metadata check output showing stale package versions

• Keeps the command output available without making the summary code-heavy.

Repro: attempted frozen uv sync command output

• Keeps the command output available without making the summary code-heavy.

Repro: attempted frozen uv export command output

• Keeps the command output available without making the summary code-heavy.

_{Ran code and verified through T-Rex}

Prompt To Fix With AI

This is a comment left during a code review.
Path: pyproject.toml
Line: 3

Comment:
**Regenerate the lockfile**

The package versions were bumped here and in `agentex/pyproject.toml`, but `uv.lock` still records both workspace packages as `0.1.0` (`agentex` and `agentex-backend`). The Docker build copies that lockfile and runs `uv sync --frozen --no-dev --package agentex-backend` and `uv export --frozen`, so the release image path can fail the frozen-lock freshness check or install/export package metadata from the previous version. Regenerate and commit `uv.lock` with the `0.1.1` package versions.

How can I resolve this? If you propose a fix, please make it concise.

[greptile-apps]

Update telemetry versions

This release updates the FastAPI/OpenAPI version to 0.1.1, but the OpenTelemetry defaults are still hard-coded to 0.1.0 in src/utils/otel_metrics.py for the SERVICE_VERSION fallback and get_meter(..., version="0.1.0"). In deployments that do not set SERVICE_VERSION, metrics emitted by the 0.1.1 service are tagged or scoped as 0.1.0, which breaks version-based dashboards and rollout comparisons. Please derive those defaults from src._version.__version__ or update them as part of the release bump.

Artifacts

Repro: Python harness that imports the real telemetry module with safe OpenTelemetry stubs and captures version defaults

• Contains supporting evidence from the run (text/x-python; charset=utf-8).

Repro: harness output showing release version 0.1.1 but telemetry service.version and meter version default to 0.1.0

• Keeps the command output available without making the summary code-heavy.

_{Ran code and verified through T-Rex}

Prompt To Fix With AI

This is a comment left during a code review.
Path: agentex/src/_version.py
Line: 3

Comment:
**Update telemetry versions**

This release updates the FastAPI/OpenAPI version to `0.1.1`, but the OpenTelemetry defaults are still hard-coded to `0.1.0` in `src/utils/otel_metrics.py` for the `SERVICE_VERSION` fallback and `get_meter(..., version="0.1.0")`. In deployments that do not set `SERVICE_VERSION`, metrics emitted by the 0.1.1 service are tagged or scoped as 0.1.0, which breaks version-based dashboards and rollout comparisons. Please derive those defaults from `src._version.__version__` or update them as part of the release bump.

How can I resolve this? If you propose a fix, please make it concise.

[github-actions]

✱ Stainless preview builds

This PR will update the agentex-sdk SDKs with the following commit messages.

openapi

chore(main): release 0.2.0

python

chore(main): release 0.2.0

typescript

chore(internal): regenerate SDK with no functional changes

Edit this comment to update them. They will appear in their respective SDK's changelogs.

✅ agentex-sdk-openapi studio · code · diff

Your SDK build had at least one "note" diagnostic, but this did not represent a regression.
generate ✅

✅ agentex-sdk-typescript studio · code · diff

Your SDK build had at least one "warning" diagnostic, but this did not represent a regression.
generate ⚠️ → build ⏭️ (prev: build ✅) → lint ⏭️ (prev: lint ✅) → test ✅

⚡ agentex-sdk-python studio · conflict

Your SDK build resulted in a merge conflict between your custom code and the newly generated changes, which is a regression from the base state.
You don't need to resolve this conflict right now, but you will need to resolve it for your changes to be released to your users. Read more about why this happened here.

---

This comment is auto-generated by GitHub Actions and is automatically kept up to date as you push.
If you push custom code to the preview branch, re-run this workflow to update the comment.
Last updated: 2026-06-24 22:59:50 UTC

[greptile-apps]

Regenerate frozen lockfile

This release bumps the workspace package to 0.2.0, but uv.lock still records both agentex and agentex-backend as 0.1.0. The Docker image copies that lockfile and runs frozen uv export / uv sync for agentex-backend, so the release build can fail the frozen-lock check or install/export package metadata for the previous version instead of this release.

Artifacts

Repro: generated frozen uv check script

• Contains supporting evidence from the run (text/x-shellscript; charset=utf-8).

Repro: frozen uv command output showing stale lockfile validation failure

• Keeps the command output available without making the summary code-heavy.

Repro: uv 0.7.3 installation output used to run the frozen checks

• Keeps the command output available without making the summary code-heavy.

_{Ran code and verified through T-Rex}

Prompt To Fix With AI

This is a comment left during a code review.
Path: pyproject.toml
Line: 3

Comment:
**Regenerate frozen lockfile**

This release bumps the workspace package to `0.2.0`, but `uv.lock` still records both `agentex` and `agentex-backend` as `0.1.0`. The Docker image copies that lockfile and runs frozen `uv export` / `uv sync` for `agentex-backend`, so the release build can fail the frozen-lock check or install/export package metadata for the previous version instead of this release.

How can I resolve this? If you propose a fix, please make it concise.

[greptile-apps]

Update telemetry version defaults

The API version now comes from __version__ and is 0.2.0, but src/utils/otel_metrics.py still falls back to SERVICE_VERSION 0.1.0 and get_meter(..., version="0.1.0"). Internal callers use get_meter without passing a version, so deployments that do not set SERVICE_VERSION emit 0.2.0 service metrics under the old 0.1.0 version, which can make rollout dashboards and version comparisons incorrect.

Artifacts

Repro: Python harness that exercises telemetry version defaults with SERVICE_VERSION unset

• Contains supporting evidence from the run (text/x-python; charset=utf-8).

Repro: harness output showing package version 0.2.0 but telemetry resource and meter version 0.1.0

• Keeps the command output available without making the summary code-heavy.

_{Ran code and verified through T-Rex}

Prompt To Fix With AI

This is a comment left during a code review.
Path: agentex/src/_version.py
Line: 3

Comment:
**Update telemetry version defaults**

The API version now comes from `__version__` and is `0.2.0`, but `src/utils/otel_metrics.py` still falls back to `SERVICE_VERSION` `0.1.0` and `get_meter(..., version="0.1.0")`. Internal callers use `get_meter` without passing a version, so deployments that do not set `SERVICE_VERSION` emit 0.2.0 service metrics under the old 0.1.0 version, which can make rollout dashboards and version comparisons incorrect.

How can I resolve this? If you propose a fix, please make it concise.