Release v1.14.0

Release v1.14.0

• agents-cli deploy runs pre-flight checks before source builds.
  • Before uploading, it checks for issues that would break the remote build and explains each one: a uv.lock pointing at a package index the builder can't reach, a missing or stale lock, build inputs excluded by .gcloudignore or .gitignore, path dependencies outside the upload, or a base image Python that doesn't match requires-python.
  • --skip-preflight skips the checks.
  • deploy --dry-run exits 1 on pre-flight errors. Scripts and CI jobs that run it may need an update.
  • #95
• agents-cli deploy adds --remove-env-vars KEY,..., --clear-env-vars, --remove-secrets KEY,... and --clear-secrets for Cloud Run and Agent Runtime.
  • A key that is in the project .env can't be removed (delete it from .env instead), and env vars agents-cli sets itself (such as AGENT_VERSION) are set again on every deploy.
  • --clear-secrets also removes secrets set outside agents-cli (such as by Terraform) and secrets mounted as volumes.
  • #60
• agents-cli deploy shows the error reason and help links when an Agent Runtime request fails.
• agents-cli extension add pypi@<package>[==<version>][#<folder>] installs an extension from a wheel on PyPI, or on the index set in UV_DEFAULT_INDEX, UV_INDEX_URL or PIP_INDEX_URL, pinned to the wheel's sha256. agents-cli extension update <name> --to <version> or --to :latest moves the pin.
• agents-cli scaffold enhance --from-adk-project writes a minimal agents-cli-manifest.yaml for an existing ADK project without rendering a template, so eval, info and playground can find it. Use --dry-run to preview and --force to replace an existing manifest.
• agents-cli scaffold create -a claude --memory-bank scaffolds a Claude agent with long-term memory in Vertex AI Memory Bank, which works the way ADK's does. The agent recalls the user's memories at the start of each turn, learns from each finished turn, and can call a search_memory tool. Other agents reject the flag; ADK agents can follow the cross-session-memory recipe.
  • Memory is per user. For A2A callers without authentication, it lasts one conversation, as in ADK.
  • With OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT on, Claude agents record the system prompt the model got, recalled memories included, as gen_ai.system_instructions on the turn span.
• Claude Agent SDK projects send x-goog-api-client: agents-cli-app/<version>+claude-agent-sdk on model calls, so Google can count usage from agents-cli projects. Set DO_NOT_TRACK=1 or AGENTS_CLI_TELEMETRY=0 to turn it off; your own ANTHROPIC_CUSTOM_HEADERS are always sent. Existing projects are unchanged until they adopt the new template code.
• agents-cli login --status, login -i and setup warn when Google Cloud credentials work but agents can't call Gemini yet, and say how to fix it.
• agents-cli login with the Gemini API key option prints the .env changes a generated project needs to use the key, instead of reporting "Continuing without authentication".
• agents-cli create explains which Google Cloud setup step failed.
• agents-cli run and agents-cli playground check the agent's Google Cloud credentials before starting it.
  • Model errors from agents-cli run, such as permission denied, disabled API, billing disabled, quota, or an invalid API key, include a one-line fix.
• agents-cli run prints the local-server banner and a failed run's error once, and shows auth errors as plain text instead of Python object reprs.
• agents-cli infra single-project stops with terraform import instructions when an existing Agent Runtime engine is missing from Terraform state, instead of creating a duplicate. On a new project the first run may log "Skipping the Agent Runtime import check" until Terraform enables the Vertex AI API.
  • #2
• google-agents-cli no longer installs scikit-learn, scipy, litellm, jsonschema or ruamel.yaml, which roughly halves its install size. Eval commands work as before.
• New ADK Python projects pin FastAPI below 0.142.0, so traces from agents on Agent Runtime and Cloud Run are no longer dropped for requests the platform forwards as unsampled.
• New Go projects use ADK Go v2.7.1.
  • For Go projects on ADK Go v2.5.0 or later, agents-cli playground honors --host and binds 127.0.0.1 by default, the same as Python projects. Pass --host 0.0.0.0 to reach it from another machine. Older projects keep binding all interfaces.
  • Action needed: when you run agents-cli scaffold upgrade on a Go project and keep your own Dockerfile, add "--host", "0.0.0.0" after "web" in its ENTRYPOINT. ADK Go v2.5.0 and later listen only on localhost by default, so the container would otherwise get no traffic.
• The setup.md prompt no longer stops after install to ask for a new session. The coding agent reads the installed workflow skill and continues.