Find code that is expensive to understand.

Read the report as a diagnostic

lean-ctx health .

The report uses structural signals such as complexity, naming and navigability. A high-cost hotspot can be worth investigating; it is not proof of a defect.

Trace the underlying source

Use the location from a report to inspect signatures, dependencies and implementation. Check whether a finding is relevant to your language, conventions and task.

Static analysis can miss dynamic relationships and can flag deliberate complexity.

Treat token tax as an estimate

A reported token or cost estimate describes the analysis model, not a measured provider bill. The amount an agent actually reads depends on its task and retrieval path.

Improve and verify

Choose a focused change, preserve behavior and run the appropriate tests. Re-run the report under the same source scope to inspect the structural difference.

Do not optimize a score at the expense of correctness or maintainability. See search and code intelligence for related tools.

Sources & versions3 references · Reviewed 17 Sept 2026

• mod.rsrust/src/core/code_health/mod.rsGitHub
• health_cmd.rsrust/src/cli/health_cmd.rsGitHub
• 22-code-health.mddocs/reference/22-code-health.mdReviewed checkout

Core checkout

0ce2207ee4

Installed runtime

3.10.2

SDK release

1.1.0

Separate baselines for source, CLI/configuration and SDK contracts. Review does not certify every platform or integration.

Versions & compatibility