Releases and Versioning
Version components
CODYGIENE reports three version numbers:
codygiene --version
codygiene 0.9.1-beta.1
engine 1.0.0
ruleset 1.0.0
- CLI version · the CLI package version (follows semver with prerelease)
- Engine version · the analysis engine version
- Ruleset version · the rule registry version
Current release
CODYGIENE 0.9.1-beta.1 is prepared for public beta testing but has not yet been published to npm. This is a prerelease (beta), not a mature stable release. Expect:
- Rule refinements based on real-world feedback
- CLI UX improvements
- Documentation updates
- Bug fixes
The frozen Ruleset 1.0.0 (44 rules) is stable: its rule IDs, categories, intended risk semantics, and ruleset identity are frozen (see below). Alpha candidate rules (26 rules) may be refined; most are enabled by default (CAND-030 and CAND-031 are disabled experimental · see below).
Versioning policy
- Major · breaking changes to CLI interface or report format
- Minor · new rules, features, or non-breaking improvements
- Patch · bug fixes and documentation updates
During the public test period, versions follow a prerelease pattern
(e.g., 0.9.1-beta.1).
Frozen Ruleset 1.0.0
The 44 frozen rules (CH-* IDs) are part of Ruleset 1.0.0. The freeze covers:
- Rule IDs · a CH-* ID keeps its meaning permanently
- Categories · a rule's category assignment does not change
- Intended risk semantics · what the rule is meant to detect
- Ruleset identity · the
ruleset 1.0.0version contract
Precision, false-positive, compatibility, and performance fixes may still be made when they preserve the intended meaning of a rule. Material semantic changes require explicit versioning · a new ruleset version, not silent edits under the same ID.
Alpha candidates
The 26 Alpha candidate rules (CAND-* IDs) may be refined based on feedback. Most are enabled by default; CAND-030 and CAND-031 are disabled experimental. If an Alpha candidate is found to produce unacceptable false positives, it may be adjusted or moved to disabled-by-default status.
Default rule profile
- Registered: 70 rules
- Enabled by default: 64 rules
- Disabled by default: 6 rules · disabled for signal-quality reasons
(high false-positive rates or unvalidated metrics), not feature gating.
Each can be enabled explicitly via
--enableor.codygiene.jsonrules.enable.
Disabled by default:
| Rule | Reason |
|---|---|
| CH-HEALTH-004 | Unused-file detection: workspace/monorepo false positives |
| CH-HEALTH-005 | Dead-code detection: workspace false positives |
| CH-HEALTH-007 | Undeclared package import: often intentional in monorepos |
| CH-HEALTH-008 | Duplicated logic: can be noisy on large codebases |
| CAND-030 | Excessive coupling: metric validation pending |
| CAND-031 | Architectural hotspot: size threshold needs validation |
Architecture insights (informational)
CODYGIENE maps workspace dependencies, identifies package cycles, and highlights modules with large architectural blast radius. These findings are marked as insights: they appear in all report formats but do not reduce quality scores or trigger release gates. Architecture insights are informational context for engineering review · not verdicts that the architecture is defective.
- CAND-029 · Change blast radius: modules with top-percentile transitive dependents
- CAND-032 · Package cycles: strongly connected workspace dependency groups
- CAND-033 · Cross-workspace dependency risk: app-to-app imports (boundary review)
CAND-030 (excessive coupling) and CAND-031 (architectural hotspot) are
disabled by default and can be enabled explicitly with --enable.
Reserved candidates
- CAND-012 · Rate limiting (reserved, not implemented)
- CAND-013 · Behavior/documentation drift (reserved, not implemented)
- CAND-014 through CAND-018 · Historical backlog reservations
Spotted something wrong or missing? Send feedback