v0.0.5

What's Changed

• PR analysis: adjust prompt to suppress LLM thinking output by @avasilevskii in #64
• PR analysis: add LLM output sanitizer by @avasilevskii in #65

Full Changelog: v0.0.4...v0.0.5