google / adk-python
Code health
This codebase scores 6.7 out of 10 for code health, which we rate fair. It also scores maintainability 7.9 and static performance 9.8 out of 10. The three are scored separately and never blended into one number. Risk is concentrated, as it usually is: 238 of 2,369 files are git hotspots, and they average 4.8 — which is where the fixes pay off most.
Do next
1 thing worth doing this quarter.
Worth planning
Delete 105 unused symbols and files (80 lines)
Nothing in the graph reaches them, across 16 files; every reader and agent pays to skip them.
- symbols
- 105
- lines
- 80
- safe to delete
- Yes(inferred)
Small effort
Improve what Repowise can see
Add a test coverage report
Without one, Repowise cannot tell tested code from untested code, so every test-related action says “unknown”.
- line coverage
- Unknown, no report
Small effortReview the 14 proposed decisions
None is accepted yet, so nothing can drift from one and agents get no enforced guidance from them.
- proposed
- 14
- accepted
- 0
Medium effort
- Lines of code
- 418.6K
Recent activity
All commitsCommits
- f4fd7d5dtest(telemetry): add property based tests for metrics export1mo ago
- 76027ddbfeat(evaluation): add optional eval set result persistence to AgentEvaluator1mo ago
- 74e7167dfix(ci): run the update-constraints hook in check mode1mo ago
- bc2c97cbfix: Key directory-loaded skill resources with forward slashes1mo ago
- 3df5a651docs: add Session unit guide1mo ago
- 461205c8feat: accept a pre-configured client on the labs OpenAI model1mo ago
Decisions
- Retain latest user input across transfer_to_agent when include_contents='none'proposed · 1mo ago
- Use a Unicode-aware tokenizer for ROUGE-1 matchingproposed · 1mo ago
- Use typing.Optional in cleanup_unused_files tool annotationsproposed · 1mo ago
- Treat RetryConfig(max_attempts=0) as no retriesproposed · 1mo ago
- Add Pydantic serializers to make session event state JSON-serializableproposed · 1mo ago
- docs(tools): encourage using mode='single_turn' in AgentTool and document manual artifacts configurationproposed · 1mo ago
Everything we found
7,648 open
Every detector's open count by area, each with its worst example. An inventory, not a to-do list: the actions above are the part worth doing.
- worst item is critical severity. Code health5,005Untested hotspot · src/google/adk/flows/llm_flows/functions.py36 findings · 10.0 deducted
- worst item is medium severity. Refactoring581build_graph+1.5 health, S effort · src/google/adk/cli/agent_graph.py
- worst item is medium severity. Decisions252tests/unittests/plugins/test_bigquery_agent_analytics_plugin.py14 awaiting review · 238 hotspots ungoverned
- worst item is medium severity. Ownership1,671src/google/adk/cli/cli_tools_click.py89% single-owner concentration, 18 commits in 90d
- worst item is low severity. Dead code139OEUnreachable function (0 lines)
Composition
Open the graphExplore this codebase
- Docs2,736 pages across 5 modules
- ChatAsk this codebase a question and get an answer with its sources.
- Files2,369 files with per-file docs, health and history
- ArchitectureDependency graph, layers, and 59 entry points
- Code healthPer-file scores, 6,303 open findings, coverage and refactoring targets
- Knowledge graphEntities, communities, and the paths between them
- Change couplingFiles that keep changing together, mined from commit history
- CommitsChange-risk ranked history with agent provenance
- ContributorsBus factor, per-file maintainers, and the human/agent split
- StatsSize class, origin, lifetime churn, rhythm and records
- CostsWhat indexing this snapshot cost, by model and by run.