v3.7.0

What's Changed

• ci: use reusable workflows for validation and fix auto-merge trigger by @CybotTM in #31
• Expand evals to 25 and improve SKILL.md workflow guidance by @CybotTM in #32

Full Changelog: v3.6.1...v3.7.0

v3.6.1

What's Changed

• fix(generator): prevent set -e exit in build_contribution_rules by @CybotTM in #29

Full Changelog: v3.6.0...v3.6.1

v3.6.0

Changes since v3.5.0

• feat: auto-create CLAUDE.md symlinks when .claude/ directory detected
• fix: trim SKILL.md to under 500 words (was 568)
• chore: add CLAUDE.md symlinks for Claude Code compatibility
• Merge pull request #26 from netresearch/feat/auto-claude-md-symlinks
• fix: always create CLAUDE.md symlinks, deprecate --no-symlinks
• Merge pull request #27 from netresearch/fix/always-create-claude-symlinks

v3.5.0

Full Changelog: v3.4.0...v3.5.0

v3.4.0

What's Changed

• chore(deps): bump the github-actions group with 3 updates by @dependabot[bot] in #24

Full Changelog: v3.3.0...v3.4.0

v3.2.0

agent-rules-skill v3.2.0

Major feature release: new extraction capabilities, domain templates, eval suite, and Agent Skills spec compliance.

New Extraction Scripts

• extract-ci-rules.sh — parses GitHub Actions/GitLab CI for quality gates, version matrix, coverage thresholds, pinned actions
• extract-architecture-rules.sh — extracts module boundaries from phpat, deptrac, Go internal/, golangci-lint depguard/forbidigo
• extract-adrs.sh — finds and summarizes architectural decision records from docs/adr/ and similar locations
• extract-github-rulesets.sh — queries GitHub rulesets API for merge queue, required checks, signed commits

Generated AGENTS.md Enhancements

• CI/Quality Gates section with version matrix and coverage thresholds
• Module Boundaries section with dependency rules and architecture patterns
• Key Decisions section with ADR summaries and file links
• Repository Settings enhanced with rulesets data

New Templates

• skill-repo.md — for Agent Skills repositories (plugin.json, SKILL.md, checkpoints)
• python-modern.md — for pyproject.toml projects with ruff, mypy, pytest
• Enhanced backend-go.md with govulncheck, fuzz testing, race detection

Infrastructure

• Agent Skills spec frontmatter (license, compatibility, metadata, allowed-tools)
• Shared validate workflow (replaced custom lint.yml)
• Version consistency gate (release blocks if tag/plugin.json/SKILL.md mismatch)
• 13 eval test cases across 3 iterations (94% pass rate, 84 assertions, +19.6% vs baseline)

Full Changelog: v3.1.0...v3.2.0

v3.1.0

What's Changed

• fix: reduce false positives in AG-02 and AG-03 checkpoints by @CybotTM in #19

Full Changelog: v3.0.2...v3.1.0

v3.0.2

What's Changed

• chore: Configure Renovate by @renovate[bot] in #16
• chore(deps): update actions/checkout action to v6 by @renovate[bot] in #17

New Contributors

• @renovate[bot] made their first contribution in #16

Full Changelog: v3.0.1...v3.0.2

v3.0.1

Full Changelog: v3.0.0...v3.0.1

v3.0.0

Breaking Changes

• Repository renamed from agents-skill to agent-rules-skill
• Composer package renamed from netresearch/agent-agents to netresearch/agent-rules-skill
• Skill directory renamed from skills/agents to skills/agent-rules
• Plugin name changed from agents to agent-rules

Why

Better reflects the broader scope: generating AGENTS.md, .github/copilot-instructions.md, and other AI agent rule files. Aligns with new convention that composer name = GitHub repo name.