With compaction.inContext enabled, compaction can generate the summary inside the cached conversation: it repeats the request the next turn would send, appends the instructions as a system message plus a user message, and sets toolChoice none. The recent tail is kept as before. A gate requires a verified model, a warm cache from the same model and thinking level, room in the context window, and a lower estimated cost than the standalone summary. Failed or rejected attempts fall back to the standalone summary, never execute tools, and are recorded in details.inContext with their usage.
pi-ai adds the supportsInContextCompaction compat setting and test/in-context-compaction-probe.ts, which measures whether a model keeps its prompt cache for this request shape. The generated catalog flags the models that passed.
Make SessionManager projections authoritative for provider requests, add append-only context edits and actionable turn boundaries, and preserve existing queue scheduling during continuation and recovery.
Run next-turn preparation only when the agent loop continues, allowing threshold compaction before the next provider request. Restore interactive progress when the same run resumes.
Closes#6879
Important to notice:
this changes when prepareNextTurn & prepareNextTurnWithContext are invoked:
the now run only when a next turn actually happens, and after shouldStopAfterTurn.
Add provider-context prefix preparation and optional summary request settings needed by cache-friendly compaction. Existing AgentSession callers continue to pass undefined, so standalone summarization remains the default until the feature is wired in.
session_before_compact and session_compact now carry
reason ("manual" | "threshold" | "overflow") and willRetry, matching
what the RPC protocol already exposes via auto_compaction_start/end.
Extensions can now distinguish manual /compact from threshold
auto-compaction and overflow recovery through the public API instead
of monkeypatching AgentSession internals.
- New compaction.md covers auto-compaction and branch summarization
- Explains cut points, split turns, data model, file tracking
- Documents session_before_compact and session_before_tree hooks
- Rewritten hooks.md matches actual API (separate event names)
- Correct ctx.ui.custom() signature (returns handle, not callback)
- Documents all session events including tree events
- Adds sessionManager and modelRegistry usage
- Updates all examples to use correct API
pi-ai:
- Fixed usage.input to exclude cached tokens for OpenAI providers
- Previously input included cached tokens, causing double-counting
- Now input + output + cacheRead + cacheWrite correctly gives total context
coding-agent:
- Session header now includes branchedFrom field for branched sessions
- Updated compaction.md with refined implementation plan
- Updated session.md with branchedFrom documentation
