v0.5.2

Patch Changes

• 9947358: Complete the AI-readable documentation contract with package-level llms and llmsFull autodiscovery fields, Claude guidance, richer workflow-agent docs, and corrected AI-skill Markdown examples.

v0.5.1

Patch Changes

• 8a90267: Add AI-readable documentation assets, llms.txt, llms-full.txt, assistant instructions, and an official Neuron-JS AI skill for coding agents.

v0.5.0

Minor Changes

• 2a6d695: Add JSON Schemas, validation helpers, normalized execution output, and explainability contracts for machine-checking generated or stored rules before runtime and auditing execution traces after a run.

v0.4.0

Minor Changes

• Align the documented public API with the package root exports.
  • Exported Synapse, ExecutionResult, lifecycle hook types, and built-in plugin classes from the package root.
  • Added AbstractAction and AbstractCondition base classes for custom extension authors.
  • Made runtime options handling safe when scripts omit optional options objects.
  • Moved tests out of src and cleaned the build before packaging to prevent stale test artifacts from reaching dist.
  • Updated implementation examples and release documentation for the expanded API surface.
• Updated GitHub Actions workflows to use Node 24-compatible official actions.
• Published via npm trusted publishing with GitHub Actions OIDC and provenance.

v0.3.0

Minor Changes

• 75526ce: First version with docs

v0.2.0

Minor Changes

• 25b6e3c: Initial modernized release of neuron-js.

  • Complete clean-room rewrite using Node 24 and Yarn 4.
  • Implemented core registry (Neuron) and execution engine (Synapse).
  • Added support for pluggable Actions, Conditions, and Parameters.
  • Built-in boolean grouping logic (AND/OR) for complex conditions.
  • Centralized lifecycle hooks system.
  • Dual-module support (ESM and CommonJS) via tshy.