fix: skip binary files with a Skipped entry; header names the directory, not its path (#21 P11)

[igorls]

Part of #21 (P11). content_filter::classify gains SkipReason::Binary (sniff via markdown::is_binary_content, same rule as the renderer); markdown::directory_display_name replaces the raw -d string in both headers and the --token-count estimate. Tests: tests/test_binary_and_header.rs (in-process classify + run_with_args + binary). One existing test (test_parallel_error_handling) expected the placeholder block; it now expects the Skipped entry.

[Copilot]

🟡 Changes recommended

Directory aliases and root paths are still labeled incorrectly, and plural binary summaries are grammatically malformed.

2 open findings

• Preserve lexical input directory basename · New
• Use plural noun for multiple binary files · New

What changed in this PR

Updates report generation to omit detected binary files and avoid exposing absolute input paths.

Changes:

• Classifies binary content as skipped and reports binary counts.
• Uses directory display names in headers and token estimates.
• Adds integration coverage and changelog documentation.

File	Description
src/​content_filter.rs	Adds binary classification and reporting.
src/​markdown.rs	Adds sanitized directory display names.
src/​lib.rs	Uses display names in token estimates.
tests/​test_binary_and_header.rs	Tests binary skipping and headers.
tests/​test_parallel_memory.rs	Updates binary-handling expectations.
CHANGELOG.md	Documents the behavior change.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.