NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
Go modules, #704 by repository stars
Latest stable
v0.2.0
Last release 1 months ago
14 Oct 2026
Ships on a steady schedule
a new release about every 9 days
Nearly every release is documented
notes for 57 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
1 years old
929 releases · first in 2025
Drag across the ruler to compare, or tap a release to open it.
security(deps): bump pytest to >=9.0.3 (CVE-2025-71176) by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/488
⚠️ This is a staging/pre-release version for testing. Not recommended for production use.
# Staging binary (use --staging flag)
curl -fsSL https://agentfield.ai/install.sh | bash -s -- --staging
# Python SDK (prerelease - requires --pre flag)
pip install --pre agentfield
# TypeScript SDK
npm install @agentfield/sdk@next
VERSION=v0.1.142-rc.5 curl -fsSL https://agentfield.ai/install.sh | bash
Download the binary for your platform below, make it executable, and move it to your PATH.
agentfield-darwin-amd64agentfield-darwin-arm64agentfield-linux-amd64agentfield-linux-arm64af call rejecting valid input for optional reasoner params by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/610_current_status issue where status stuck on `S… by @DebanKsahu in https://github.com/Agent-Field/agentfield/pull/673af install by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/738Note truncated.
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Bumps the npm_and_yarn group with 1 update in the /desktop directory: electron. Bumps the npm_and_yarn group with 1 update in the /sdk/typescript directory: axios.
Updates electron from 42.9.1 to 42.10.0
Updates axios from 1.18.0 to 1.20.0
updated-dependencies:
Signed-off-by: dependabot[bot] support@github.com Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> (484bddf)
security(deps): bump pytest to >=9.0.3 (CVE-2025-71176) by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/488
⚠️ This is a staging/pre-release version for testing. Not recommended for production use.
# Staging binary (use --staging flag)
curl -fsSL https://agentfield.ai/install.sh | bash -s -- --staging
# Python SDK (prerelease - requires --pre flag)
pip install --pre agentfield
# TypeScript SDK
npm install @agentfield/sdk@next
VERSION=v0.1.142-rc.4 curl -fsSL https://agentfield.ai/install.sh | bash
Download the binary for your platform below, make it executable, and move it to your PATH.
agentfield-darwin-amd64agentfield-darwin-arm64agentfield-linux-amd64agentfield-linux-arm64af call rejecting valid input for optional reasoner params by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/610_current_status issue where status stuck on `S… by @DebanKsahu in https://github.com/Agent-Field/agentfield/pull/673af install by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/738Note truncated.
security(deps): bump pytest to >=9.0.3 (CVE-2025-71176) by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/488
⚠️ This is a staging/pre-release version for testing. Not recommended for production use.
# Staging binary (use --staging flag)
curl -fsSL https://agentfield.ai/install.sh | bash -s -- --staging
# Python SDK (prerelease - requires --pre flag)
pip install --pre agentfield
# TypeScript SDK
npm install @agentfield/sdk@next
VERSION=v0.1.142-rc.3 curl -fsSL https://agentfield.ai/install.sh | bash
Download the binary for your platform below, make it executable, and move it to your PATH.
agentfield-darwin-amd64agentfield-darwin-arm64agentfield-linux-amd64agentfield-linux-arm64af call rejecting valid input for optional reasoner params by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/610_current_status issue where status stuck on `S… by @DebanKsahu in https://github.com/Agent-Field/agentfield/pull/673af install by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/738Note truncated.
Fix(control-plane): return 404 for status callbacks on unknown executions (#1080) (e447e0b)
Fix(sdk/python): parse AGENTFIELD_ASYNC boolean flags with the repo's env vocabulary (#1082)
AsyncConfig.from_environment() read all five boolean feature flags with
lambda x: x.lower() == "true". Four of those fields default to True, so the
only values that could be expressed were "true" and "false": setting
AGENTFIELD_ASYNC_ENABLE_RESULT_CACHING=1 -- or =yes, or =on -- evaluated to
False and silently disabled result caching, the opposite of what the value says.
The same inversion applied to enable_async_execution, enable_batch_polling and
fallback_to_sync, and it reaches users by default because both Agent (agent.py:846)
and AgentFieldClient (client.py:182) build their async_config from
from_environment() when none is passed.
The lambda also never raises, so get_env_var's except (ValueError, TypeError): return default_value fallback is unreachable for booleans and an unparseable
value overrides the field with False instead of leaving the default. PR #714,
which wired this into the client default, documents the intended contract:
from_environment() "only overrides fields when the corresponding env var is set
(falling back to the default on unparseable values)".
Use the vocabulary the package already uses, split by the field's default: default-on flags opt out with ("0", "false", "no", "off") like log_writer._queue_enabled, logger._stdout_mirror_enabled, node_logs.logs_enabled and openrouter_attribution.attribution_enabled; the default-off enable_event_stream opts in with ("1", "true", "yes", "on") like litellm_observability._TRUE_VALUES. A value outside the vocabulary now leaves the field at its default, which is what makes the documented fallback hold without the converter having to raise.
With no AGENTFIELD_ASYNC_* variable set the converters are never reached -- get_env_var returns the default before consulting them -- so from_environment() produces a byte-identical AsyncConfig to before (verified by digesting every field on both trees). Only explicit opt-in/opt-out values change behaviour, and every value that already worked still does: "true"/"false"/"0" are unchanged, while "1"/"yes"/"on"/whitespace-padded values now mean what they say.
logger.py's three AGENTFIELD_LOG_* flags use the same == "true" shape but all
default to False, so they fail in the less harmful direction; left alone here to
keep this to one seam. (e960677)
security(deps): bump pytest to >=9.0.3 (CVE-2025-71176) by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/488
⚠️ This is a staging/pre-release version for testing. Not recommended for production use.
# Staging binary (use --staging flag)
curl -fsSL https://agentfield.ai/install.sh | bash -s -- --staging
# Python SDK (prerelease - requires --pre flag)
pip install --pre agentfield
# TypeScript SDK
npm install @agentfield/sdk@next
VERSION=v0.1.142-rc.2 curl -fsSL https://agentfield.ai/install.sh | bash
Download the binary for your platform below, make it executable, and move it to your PATH.
agentfield-darwin-amd64agentfield-darwin-arm64agentfield-linux-amd64agentfield-linux-arm64af call rejecting valid input for optional reasoner params by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/610_current_status issue where status stuck on `S… by @DebanKsahu in https://github.com/Agent-Field/agentfield/pull/673af install by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/738Note truncated.
_calculate_backoff_delay called random.seed(self._container_seed + attempt) before drawing jitter, which reseeds the module-level generator shared by the whole host process. Every rate-limit retry therefore reset the application's random stream to a value derived from HOSTNAME and the pid, so all later random.random()/choice()/shuffle()/sample() calls in the process returned a repeatable sequence instead of continuing the stream the caller seeded.
It also degraded the SDK's own jitter: rate_limiter.py is the only module that seeds the global generator, but client.py:1295 (_next_poll_interval) and harness/_runner.py:415 and :423 (transient-error backoff) draw from it, so one rate-limited LLM call made their poll intervals and backoff delays a deterministic function of the container seed -- identical across containers that share a hostname/pid shape, which is exactly the synchronisation jitter exists to prevent.
Draw from a private random.Random instance instead. The returned jitter is bit-identical (verified across 240 seed/attempt/range combinations by comparing float.hex()), because CPython's module-level random.uniform delegates to a module-level Random seeded through the same path. Backoff timing, retry counts, and per-container jitter distribution are unchanged; the seed just stops leaking into global state, which also stops concurrent callers from drawing jitter off a seed another caller installed.
This matches what the Go and TypeScript SDKs already do: sdk/go/ai/ rate_limiter.go:321 uses rand.New(rand.NewSource(containerSeed + attempt)) and sdk/typescript/src/ai/RateLimiter.ts:172 uses a local _createJitterRng closure, both with the same seeding scheme and the same (u*2-1)*r formula that random.uniform(-r, r) expands to.
The existing test_calculate_backoff_applies_jitter_and_max_cap already computes its expectation with random.Random(container_seed + attempt) and compares only the returned number, which is why this survived: nothing asserted on the global state. Adds two tests pinning the contract at both the helper and the public execute_with_retry path.
enable_rate_limit_retry defaults to True (types.py:497) and agent_ai.py:605-609 re-forces it, so the affected path is the default one. (b675e2d)
security(deps): bump pytest to >=9.0.3 (CVE-2025-71176) by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/488
⚠️ This is a staging/pre-release version for testing. Not recommended for production use.
# Staging binary (use --staging flag)
curl -fsSL https://agentfield.ai/install.sh | bash -s -- --staging
# Python SDK (prerelease - requires --pre flag)
pip install --pre agentfield
# TypeScript SDK
npm install @agentfield/sdk@next
VERSION=v0.1.142-rc.1 curl -fsSL https://agentfield.ai/install.sh | bash
Download the binary for your platform below, make it executable, and move it to your PATH.
agentfield-darwin-amd64agentfield-darwin-arm64agentfield-linux-amd64agentfield-linux-arm64af call rejecting valid input for optional reasoner params by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/610_current_status issue where status stuck on `S… by @DebanKsahu in https://github.com/Agent-Field/agentfield/pull/673af install by @AbirAbbas in https://github.com/Agent-Field/agentfield/pull/738Note truncated.
Bumps the uv group with 1 update in the /sdk/python directory: pyjwt.
Updates pyjwt from 2.13.0 to 2.15.0
updated-dependencies:
Signed-off-by: dependabot[bot] support@github.com Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> (88ae7ca)
Bumps the npm_and_yarn group with 1 update in the /examples/benchmarks/100k-scale/mastra-bench directory: ip-address.
Updates ip-address from 10.4.0 to 10.7.2
updated-dependencies:
Signed-off-by: dependabot[bot] support@github.com Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> (c5750ee)
`bash curl -fsSL https://agentfield.ai/install.sh | bash `
curl -fsSL https://agentfield.ai/install.sh | bash
VERSION=v0.1.141 curl -fsSL https://agentfield.ai/install.sh | bash
Download the binary for your platform below, make it executable, and move it to your PATH.
agentfield-darwin-amd64agentfield-darwin-arm64agentfield-linux-amd64agentfield-linux-arm64Full Changelog: https://github.com/Agent-Field/agentfield/compare/v0.1.140...v0.1.141
One column per month.