getcompanion-ai/co-mono
1
0
Fork 0
mirror of https://github.com/getcompanion-ai/co-mono.git synced 2026-04-18 19:03:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Simplify compaction: remove proactive abort, use Agent.continue() for retry

Browse source 
- Add agentLoopContinue() to pi-ai for resuming from existing context
- Add Agent.continue() method and transport.continue() interface
- Simplify AgentSession compaction to two cases: overflow (auto-retry) and threshold (no retry)
- Remove proactive mid-turn compaction abort
- Merge turn prefix summary into main summary
- Add isCompacting property to AgentSession and RPC state
- Block input during compaction in interactive mode
- Show compaction count on session resume
- Rename RPC.md to rpc.md for consistency

Related to #128
This commit is contained in:
Mario Zechner 2025-12-09 21:43:49 +01:00
parent d67c69c6e9
commit 5a9d844f9a
27 changed files with 1261 additions and 1011 deletions
Download patch file Download diff file Expand all files Collapse all files

6
packages/coding-agent/README.md
View file

@ -267,7 +267,11 @@ Long sessions can exhaust context windows. Compaction summarizes older messages
**Manual:** `/compact` or `/compact Focus on the API changes`
**Automatic:** Enable with `/autocompact`. Triggers when context exceeds threshold.
**Automatic:** Enable with `/autocompact`. When enabled, triggers in two cases:
- **Overflow recovery**: LLM returns context overflow error. Compacts and auto-retries.
- **Threshold maintenance**: Context exceeds `contextWindow - reserveTokens` after a successful turn. Compacts without retry.
When disabled, neither case triggers automatic compaction (use `/compact` manually if needed).
**How it works:**
1. Cut point calculated to keep ~20k tokens of recent messages