Skip to content

v2.7.0

Latest
Latest

Choose a tag to compare

View all tags
@wolo-lab wolo-lab released this 02 Nov 10:37
· 1 commit to main since this release
v2.7.0
be8b100
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Highlights

v2.7.0 adds Anthropic Claude as a first-party model, search-gated tool discovery for large tool catalogs, an experimental classifier API for workflow routing, and a per-invocation budget on model calls. It also switches telemetry to the OpenTelemetry GenAI semantic conventions by default, which is a breaking change for anyone consuming ADK traces.

  • Anthropic Claude models (experimental): model/anthropicmodel implements model.LLM over the Anthropic Messages API and Vertex AI, with streaming, tool calling, image and PDF input, extended and adaptive thinking, structured output and prompt caching. See examples/anthropic. (#1866)
  • Search-gated tool discovery: tool/toolsearch wraps any tool.Toolset so the model starts with a search_tools tool plus a configured core set, and the tools it finds are declared from the next model step on. This keeps a large MCP catalog out of every request. See examples/toolsearch. (#1734, #1735)
  • Classifiers (experimental): exp/classifier defines a Classifier that answers typed Choice, Binary and Score questions with probabilities. It ships with a client for TypeSafe's Jev decision models (exp/classifier/typesafe), an LLMClassifier backed by any model.LLM, and a workflow Node that routes on the answer, matching adk-python's ClassifierNode. (#2003, #2005, #2035)
  • Model-call budget: RunConfig.MaxLLMCalls bounds the model calls one invocation may make, 500 by default, and ends the run with agent.ErrLLMCallsLimitExceeded when the budget is spent. Matches adk-python's max_llm_calls. This changes behavior for existing users, see below. (#1412)
  • Telemetry: traces follow OpenTelemetry GenAI semantic conventions v2 by default, as adk-python does. This is breaking for telemetry consumers, see below. (#1635)
  • OpenAI-compatible models: images and files in InlineData and FileData parts are sent on both the Responses and Chat Completions endpoints (#1543). A response schema title that is not a valid format name falls back to adk_response instead of a 400 from the API (#1912). A GenerateContentConfig field that a newer genai adds is refused with ErrUnsupportedConfigField instead of being dropped silently (#1922).
  • Gemini built-in tools: geminitool.GoogleMaps, geminitool.UrlContext and geminitool.EnterpriseWebSearch. (#2026)
  • MCP: with Config.Auth set, each user gets their own MCP session, so servers that bind a session to the user who opened it stop answering other users with 403 (#1755). MCP HTTP requests identify ADK in User-Agent and X-Goog-Api-Client, as model calls already do (#1862).
  • Auth fix: credentials from auth.ADC() and auth.ServiceAccount() failed every token mint with context canceled on local gcloud ADC and impersonated service accounts. Services on GCE or with a key file were not affected. (#1904)
  • Tool auth groundwork: tool/toolauth adds adk-python's auth data model and wire format for adk_request_credential, so requests, replies and credential keys agree with adk-python and ADK Web. Nothing uses it yet. (#2032)
  • Reliability:
    • Workflow joins resume with completed predecessors restored instead of rerunning them. (#1572)
    • LoopAgent stops on context cancellation. WithAgentTimeout and WithAgentCancel work on tool and callback contexts instead of returning nil. (#1810, #1604)
    • Artifacts saved in agent and model callbacks are recorded in ArtifactDelta. OutputKey is saved when a before-agent callback short-circuits. (#1878, #1800, #1963)
    • Optional instruction placeholders such as {key?} and {artifact.name?} surface infrastructure errors instead of rendering an empty string. This changes behavior, see below. (#1748)
    • Vertex AI sessions persist live transcriptions. In-memory memory search skips thought parts. (#1964, #1975)
    • Streaming function tools preserve consumer panics. agenttool keeps code-execution output in sub-agent results. Running an agent without Run returns an error instead of doing nothing. (#1961, #1860, #1850)
    • launcher.Config.TelemetryOptions honors WithOtelToCloud. Live reconnect matches close code 1008 by type, not by error text. (#1968, #1993)
  • CLI: adkgo deploy agentengine --mem_deploy no longer defaults Memory Bank to gemini-2.5-flash, which Vertex AI stopped serving to new projects. Memory Bank applies its own default model unless --mem_model is set. (#1902)
  • Agent Analytics plugin (plugin/agentanalytics, its own module and tag): aligned with adk-python's schema and event vocabulary, with rich session, HITL, A2A and workflow events, GCS offloading of multimodal content, opt-in exactly-once delivery, drop statistics, hierarchical span ids with latency and time-to-first-token, and 25 pre-built per-event BigQuery views. (#1786, #1787, #1788, #1789, #1790, #1791 and follow-up fixes)

Breaking changes

  • Telemetry schema: ADK now emits the OpenTelemetry GenAI semantic conventions v2. The legacy log events are replaced by one gen_ai.client.inference.operation.details event per model call, message content on spans is a structured value rather than a JSON string, execute_tool records arguments and results only when span content capture is on (under gen_ai.tool.call.*), and an LLM-agent root gets a new invoke_workflow parent span. Set ADK_TELEMETRY_SCHEMA_VERSION_OPT_IN=otel_semconv_1_36 to restore the previous format, supported until at least March 31, 2027. (#1635)
  • Model calls are capped at 500 per invocation: a run that makes more than 500 model calls now ends with an error wrapping agent.ErrLLMCallsLimitExceeded. Raise the cap with RunConfig.MaxLLMCalls or ADK_MAX_LLM_CALLS, or set MaxLLMCalls to -1 for no limit. Unlike adk-python, 0 means the default, not unlimited. (#1412)
  • Optional instruction placeholders propagate errors: {key?} and {artifact.name?} render empty only when the key or artifact is missing (session.ErrStateKeyNotExist, fs.ErrNotExist). Any other error from the session or artifact service now fails the run. A custom artifact.Service must wrap fs.ErrNotExist for missing files, and an agenttool sub-agent with an optional artifact placeholder now errors when the root runner has no artifact service. (#1748)
  • Go 1.26.10 is required: go.mod moves from 1.26.6 to 1.26.10. (#1916)
  • Agent Analytics plugin defaults: the default table name is agent_events instead of events, existing tables are upgraded additively on startup unless DisableSchemaUpgrade is set, the per-event views are created on startup unless DisableCreateViews is set, and granular event types (STATE_DELTA, AGENT_TRANSFER, HITL_*, NODE_OUTPUT, …) replace the generic EVENT row. These apply when the plugin module is next tagged, not to the core v2.7.0 tag. (#1786, #1788, #1791)

Smaller behavior changes:

  • geminitool.GoogleSearch returns an error on a non-Gemini model, and on a Gemini 1.x model combined with other tools. (#2026)
  • launcher: passing telemetry.WithOtelToCloud(true) in TelemetryOptions now exports to Google Cloud and needs credentials at startup, where it used to start with export silently off. (#1968)
  • openaimodel: a config field added by a genai release newer than v1.74.0 is refused with ErrUnsupportedConfigField. Configs valid today are unaffected. (#1922)
  • mcptoolset with Config.Auth opens one session per distinct credential, closes sessions idle for 10 minutes, and caps the pool at 64. Without Auth nothing changes. (#1755)
  • LoopAgent yields the context error once and stops when its context is cancelled, where it used to keep looping if no pass yielded an event. (#1810)
  • An agent callback that saves an artifact and returns nil now yields one event carrying the ArtifactDelta, where it yielded none. (#1878)
  • The in-memory memory service no longer stores or matches thought parts, and an event with only thought text is not stored. (#1975)
  • Deprecated: artifact.SaveRequest.Version. Save has never read it. (#1985)
All changes

What's Changed

  • feat(tool): add search-gated tool discovery by @wolo-lab in #1734
  • docs(examples/toolsearch): add a tool search sample by @wolo-lab in #1735
  • fix(agent): report an error when an agent without Run is run by @wolo-lab in #1850
  • docs(artifact): fix service references in doc comments by @ktsoator in #1775
  • feat(agentanalytics): align schema, event vocabulary, redaction, and core contract with adk-python by @haiyuan-eng-google in #1786
  • fix(loopagent): stop on context cancellation by @anzal1 in #1810
  • feat(runner): bound model calls per invocation with RunConfig.MaxLLMCalls by @venkat-uk in #1412
  • fix(llminternal): prevent silent context degradation by propagating systemic state failures by @karthik-0306 in #1748
  • test(model/gemini): pin the last citation and accepting AutomaticContinuation by @baptmont in #1825
  • test(llminternal): match isResumable's doc and test table to the errors it sees by @harshitwandhare in #1750
  • docs(loopagent): say the context is checked before each pass by @anzal1 in #1861
  • refactor(workflow): remove redundant goroutine in scheduleRetry by @moodbased in #1849
  • docs(compaction): list the GenerateContentConfig fields the summarizer uses by @karolpiotrowicz in #1828
  • fix(agenttool): preserve code output in sub-agent results by @ktsoator in #1860
  • feat(mcptoolset): identify ADK in the User-Agent of MCP HTTP requests by @wolo-lab in #1862
  • fix(mcptoolset): stop using context.Background() in connectionRefresher.Close by @wolo-lab in #1844
  • feat(model/openaimodel): support image/file input via genai InlineData/FileData parts by @tonydzi in #1543
  • feat(agentanalytics): add concurrency-safe trace manager, latency_ms, and streaming TTFT by @haiyuan-eng-google in #1787
  • fix(llminternal): record model callback artifact saves on the event by @Vivek1106-04 in #1800
  • fix(agent): record artifact-only deltas from before/after agent callbacks by @anzal1 in #1878
  • fix(workflow): restore completed predecessors when resuming joins by @lucky798213 in #1572
  • fix(agent): make WithAgentTimeout work on tool and callback contexts by @jaideeppyne in #1604
  • test(model/gemini): pin that MAX_TOKENS after CONTINUATION is not resumed by @beiryu in #1872
  • chore(deps): bump the go-minor-patch group with 2 updates by @dependabot[bot] in #1901
  • chore(deps): bump golang.org/x/net to v0.60.0 in agentanalytics by @wojtas000 in #1908
  • feat(telemetry)!: default to experimental GenAI semconv, with ADK_TELEMETRY_SCHEMA_VERSION_OPT_IN=otel_semconv_1_36 rollback by @RKest in #1635
  • fix(adkgo/deploy): stop defaulting Memory Bank to gemini-2.5-flash by @wojtas000 in #1902
  • feat(agentanalytics): emit rich session/HITL/A2A/workflow events and classify tool_origin by @haiyuan-eng-google in #1788
  • Upgrading go to 1.26.10 by @kdroste-google in #1916
  • fix(agentanalytics): redact credentials when a value ends in a key name by @haiyuan-eng-google in #1863
  • feat(agentanalytics): add race-free GCS multimodal offloading and ObjectRef support by @haiyuan-eng-google in #1789
  • chore(v2): update Go version requirements in documentation by @ktsoator in #1926
  • fix(agentanalytics): preserve live spans when tasks resume by @jabrailkhalil in #1910
  • fix(agentanalytics): never mark workflow agents taskDelegated by @haiyuan-eng-google in #1935
  • feat(agentanalytics): pause GCS uploads after batch timeout and preserve run-end flush budget by @haiyuan-eng-google in #1936
  • feat(agentanalytics): add DropStats, ExactlyOnceDelivery, and schema setup hardening by @haiyuan-eng-google in #1790
  • fix(agentanalytics): preserve invocation spans for late callbacks by @jabrailkhalil in #1911
  • fix(agentanalytics): ExactlyOnceDelivery and DropStats follow-ups from #1790 by @haiyuan-eng-google in #1943
  • feat(agentanalytics): add pre-built per-event BigQuery analytics views by @haiyuan-eng-google in #1791
  • chore(deps): bump the go-minor-patch group with 3 updates by @dependabot[bot] in #1969
  • fix(agentanalytics): log a deferred GCS retry only when it is requeued by @leilei3167 in #1941
  • fix(session/vertexai): persist live transcriptions by @DisturbedCrow in #1964
  • fix(agentanalytics): pause stalled run-end uploads across batches by @ktsoator in #1947
  • fix(llmagent): save output_key when a before-agent callback short-circuits by @anzal1 in #1963
  • fix(functiontool): preserve streaming consumer panics by @jabrailkhalil in #1961
  • fix(launcher): honor WithOtelToCloud in launcher.Config.TelemetryOptions by @Saurav-Gupta-9741 in #1968
  • fix(memory): skip thought parts in in-memory search and preloaded memories by @Vivek1106-04 in #1975
  • chore(deps): bump the go_modules group across 1 directory with 2 updates by @dependabot[bot] in #1934
  • fix(auth): keep the init context alive so token sources can refresh by @wojtas000 in #1904
  • fix(openaimodel): use adk_response when the schema title is not a valid format name by @wojtas000 in #1912
  • fix(openaimodel): refuse config fields genai adds before they are supported by @wojtas000 in #1922
  • fix(ci): stop every CI run from building with an empty Go cache by @baptmont in #1972
  • fix(agentanalytics): report the right terminal row in more failure cases by @haiyuan-eng-google in #1952
  • feat(model/anthropicmodel): add Anthropic Claude model support by @ScottMansfield in #1866
  • docs(artifact): deprecate SaveRequest.Version by @wojtas000 in #1985
  • fix(agentanalytics): cap deferred retry upload budget and pause on shorter run-end stalls by @haiyuan-eng-google in #1955
  • fix(llminternal): match close code 1008 by type in isResumable by @wojtas000 in #1993
  • fix(agentanalytics): tidy span bookkeeping, comments and tests by @haiyuan-eng-google in #1953
  • fix(agentanalytics): release deferred upload reservations on append or Close panic by @haiyuan-eng-google in #1954
  • fix(mcptoolset): give each user its own MCP session with Config.Auth by @Vivek1106-04 in #1755
  • feat(exp): add classifier interface and TypeSafe Jev decision model client by @rafal-glowinski in #2003
  • Added google_maps, url_contex + gemini guards by @kdroste-google in #2026
  • feat(exp): add workflow routing Node for classifiers by @rafal-glowinski in #2005
  • feat(exp): add LLMClassifier backed by model.LLM by @rafal-glowinski in #2035
  • feat(toolauth): add adk-python's auth data model and wire format by @jjsasha63 in #2032
  • Update version.go by @baptmont in #2031

New Contributors

Full Changelog: v2.6.0...v2.7.0

New Contributors

Contributors

ScottMansfield, rafal-glowinski, and 24 other contributors
Assets 2
Loading