v0.5.0 — AI Query Interface (query + cache CLI)

What's New

intentgraph query — 7 sub-commands

Machine-readable CLI interface for AI agents to query cached repository analysis:

• query symbols <file> — list all symbols in a file
• query context <file> — full context: language, loc, sha256, symbols, exports, deps, dependents
• query deps <file> — direct dependencies
• query dependents <file> — files that depend on a given file
• query callers <symbol> — call sites for a symbol
• query path <file-a> <file-b> — shortest dependency path (BFS)
• query search — filter files by name pattern, language, symbol presence, complexity

intentgraph cache — 3 sub-commands

Analysis cache management (no re-analysis on every query):

• cache status — existence, staleness, file count, cache path
• cache warm — pre-warm cache (or use existing)
• cache clear — delete cached analysis

Cache behaviour

• Stored at <repo>/.intentgraph/cache.json
• Atomic writes (tempfile + os.replace)
• SHA-256 staleness detection per file
• schema_version=\"1\" guard prevents stale deserialization

Test coverage

125/125 unit + CLI tests passing. Integration-tested against the intentgraph repo itself (45 Python files, 24/24 CLI commands verified).

Full changelog

See PR #6 for detailed commit history.