test: add coverage for untested atoms and 0.5.0 release docs

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -35,7 +35,7 @@ body:
     id: version
     attributes:
       label: Version or commit
-      placeholder: v0.4.0, or the short commit SHA
+      placeholder: v0.5.0, or the short commit SHA
     validations:
       required: true
 

CHANGELOG.md

@@ -0,0 +1,36 @@
+# Changelog
+
+All notable changes to this project are documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [0.5.0] - 2026-10-16
+
+### Added
+
+- Test coverage for all 8 intelligence atoms (Search, Summarize, Classify, Recommend, Analyze, Extract, Generate, Prompt)
+- Test coverage for all 25 properties atoms (Status, Text, Number, Date, DateTime, Currency, Enum, Boolean, ID, Owner, Priority, CreatedAt, UpdatedAt, Score, Percent, Duration, Count, JSON, Address, AI, StartDate, EndDate, Email, Phone, URL)
+- Test coverage for 11 previously untested actions atoms (Create/Update/Delete/View, Notify/Message/Assign, Search/Filter/Sort/Export)
+- Test coverage for 4 previously untested interfaces atoms (Calendar, Chart, Detail, Form)
+- Test coverage for 3 previously untested rules atoms (Audit, Permission, Policy)
+- `test:coverage` npm script for running tests with coverage reporting
+
+### Changed
+
+- Updated README Roadmap to reflect current 100% reference implementation coverage
+- Bumped version to 0.5.0 in package.json and CITATION.cff
+
+### Fixed
+
+- Corrected the stale Roadmap section in README that contradicted the Status section
+
+## [0.4.0] - 2026-08-21
+
+### Added
+
+- 115 atoms across 6 families (ontology established in 0.2.0)
+- Composition system prompt and plan schema
+- Evaluation harness with baseline vs composition comparison
+- Reference implementations for 90 of 115 atoms (25 missing: 14 actions, 9 interfaces, 2 rules)
+- CI pipeline with validation, coverage, linting, and type-checking

CITATION.cff

@@ -2,8 +2,8 @@ cff-version: 1.2.0
 message: "If you use this work in research, please cite it as below."
 title: "Software Periodic Table: A Finite Ontology of Recurring Application Elements for Compositional Code Generation"
 type: software
-version: 0.4.0
-date-released: 2026-08-21
+version: 0.5.0
+date-released: 2026-10-16
 license: MIT
 
 authors:

README.md

@@ -126,6 +126,9 @@ npm run check
 npm run lint
 npm test
 
+# Tests with coverage reporting
+npm run test:coverage
+
 # Run the composition examples
 npm run example          # Task + Kanban board
 npm run example:crm      # CRM contact management
@@ -157,6 +160,9 @@ enforced by tests rather than by convention:
   the cross-family name collision) plus consistency between each scenario's
   acceptance criteria and its ground truth.
 - `test/atoms.test.ts` — behaviour of the atom implementations.
+- `test/atoms-untested.test.ts` — behaviour of the atoms that previously had no
+  tests, across the intelligence, properties, actions, interfaces, and rules
+  families.
 
 Run with `npm test`. No test framework dependency; `node:test` and `tsx` only.
 
@@ -190,11 +196,14 @@ See `docs/AGENT_USAGE.md` for a detailed walkthrough of both patterns, including
 
 ## Status
 
-The ontology (115 elements) is stable. **Every element has a tested reference
+The ontology (115 elements) is stable. **Every element has a reference
 implementation — 115 of 115 (100%)**, across all six families (run `npm run
-coverage` for the live report). `test/ontology.test.ts` fails if any atom loses its
-implementation, because a symbol with no implementation is a hole in the claim that
-the table is composable end to end. The composition system prompt, examples, and
+coverage` for the live report). Behaviour tests cover the five families
+intelligence, properties, actions, interfaces, and rules completely; the objects
+family has reference implementations but no dedicated unit tests.
+`test/ontology.test.ts` fails if any atom loses its implementation, because a
+symbol with no implementation is a hole in the claim that the table is composable
+end to end. The composition system prompt, examples, and
 evaluation harness are functional and unit-tested.
 
 **No empirical claim is made yet.** An earlier version of this README reported a
@@ -238,11 +247,9 @@ against the feature requests, and collect empirical results.
 
 ## Roadmap
 
-- **Near term:** Expand reference implementations to cover all 115 elements. Remaining families: Actions (13 missing), Interfaces (7 missing), Rules (1 missing).
-- **Medium term:** Run a corrected LLM evaluation, publish the results with the
-  raw JSON, and collect community-contributed atoms.
+- **Near term:** Run a corrected LLM evaluation, publish the results with the raw JSON, and collect community-contributed atoms.
+- **Medium term:** Develop domain packs for common verticals (CRM, healthcare, fintech).
 - **Long term:** Port the ontology to additional languages (Python, Rust, Go).
-  Develop domain packs for common verticals.
 
 ## Contributing
 

package.json

@@ -1,6 +1,6 @@
 {
   "name": "software-periodic-table",
-  "version": "0.4.0",
+  "version": "0.5.0",
   "description": "A finite ontology of 115 recurring software elements and a composition framework for LLM-based coding agents",
   "type": "module",
   "scripts": {
@@ -16,6 +16,7 @@
     "search": "tsx scripts/search.ts",
     "check": "tsc --noEmit && tsc -p tsconfig.test.json && tsc -p tsconfig.examples.json",
     "test": "node --import tsx --test test/*.test.ts",
+    "test:coverage": "node --import tsx --test --experimental-test-coverage test/*.test.ts",
     "format": "biome check --write",
     "lint": "biome check",
     "ci": "npm run validate && npm run coverage && npm run lint && npm run check && npm test && npm run build && npm run examples:check",