Release v0.10.0: evaluation framework

New commands:
  agent-strace eval run <session-id>           score a session
  agent-strace eval compare <session-a> <b>    side-by-side diff
  agent-strace eval ci <session-id>            CI gate (exits 1 on fail)
  agent-strace eval dataset add|list|export    manage eval datasets

Built-in scorers: no_errors, regex, cost_under, files_scoped,
duration_under, custom. Config via .agent-evals.yaml (stdlib YAML
parser, no PyYAML). Dataset storage is local JSONL.

Zero new dependencies. 47 new tests.

Co-authored-by: Ona <no-reply@ona.com>

Author: Siddhant-K-code

src/agent_trace/__init__.py

@@ -1,3 +1,3 @@
 """agent-trace: strace for AI agents."""
 
-__version__ = "0.9.0"
+__version__ = "0.10.0"