Feat: add agent run cache

[declanjackson]

Add Agent Run Cache

• Add caching and resumption functionality for agent tasks
• Automatically cache agent state (conversation history, execution environment files) on interruption (Ctrl+C, error, or max turns reached)
• Enable resumption via resume=True parameter on session() - if a cache exists for the same prompt, the agent restores state and continues from the last turn
• Cache is keyed by SHA256 hash of initial messages and stored in ~/.cache/stirrup/<task_hash>/

How to test/use:

Enable resumption: Pass resume=True to session():

async with agent.session(output_dir="./output", resume=True) as session:
     await session.run("Analyze all datasets")

Preserve cache after success: Set clear_cache_on_success=False to keep cache for debugging:

async with agent.session(resume=True, clear_cache_on_success=False) as session:
    await session.run("Long running task")

Files changed

File	Change
src/stirrup/core/cache.py	New - Cache module with CacheState, CacheManager, and serialization helpers
src/stirrup/core/agent.py	Modified - Added resume/cache logic, signal handling for graceful interruption
src/stirrup/tools/code_backends/base.py	Modified - Added temp_dir property to base class
docs/guides/caching.md	New - Documentation for the caching feature
mkdocs.yml	Modified - Added caching guide to navigation