v0.2.0

Homer v0.2.0 — First Tagged Release

Repository mining tool for agentic development context.

Pipeline

• Extract → Auto Snapshots → Analyze → Render with topological ordering
• Incremental: checkpoint-based needs_rerun() skips unchanged analysis (544ms no-op vs 12s full)
• 7 extractors: git, structure, graph, document, github, gitlab, prompt
• 7 analyzers: behavioral, centrality, community, temporal, convention, task_pattern, semantic
• 6 renderers: agents-md, module-ctx, risk-map, skills, topos-spec, report

Language Support

13 languages with precise scope graph resolution via tree-sitter:
Rust, Python, TypeScript, JavaScript, Go, Java, Ruby, Swift, Kotlin, C#, PHP, Zig, Lean 4

CLI

10 commands: init, update, status, query, graph, diff, render, snapshot, risk-check, serve

• risk-check --diff <base-ref> for PR-scoped risk assessment
• diff command with impact analysis (salience, bus factor, community mapping)

MCP Server

6 tools for AI agent integration: homer_query, homer_graph, homer_risk, homer_co_changes, homer_conventions, homer_diff

Output Quality

• Semantic community labels in AGENTS.md (16 pattern mappings with deduplication)
• Staleness indicators in all artifacts
• Module context files (.context.md) for directory-level summaries

Quality

• 522 tests (including 42 property tests via proptest)
• Clippy pedantic clean (-D warnings)
• Edition 2024, MSRV 1.85, unsafe forbidden