-
-
Notifications
You must be signed in to change notification settings - Fork 1
Research findings are paid for and thrown away — persist them as reusable context chunks #1963
Copy link
Copy link
Open
Labels
P2Polish — worth doing, not urgentPolish — worth doing, not urgentarea:contextstella-context — graph, embeddings, memory, retrievalstella-context — graph, embeddings, memory, retrievalarea:pipelinestella-pipeline — triage→…→judge orchestrationstella-pipeline — triage→…→judge orchestrationfeatureNew capability or improvementNew capability or improvementgoal:enginePillar 4 — engine tuned and optimized: budget, wall-clock, resolve ratePillar 4 — engine tuned and optimized: budget, wall-clock, resolve ratepain:token-efficiencyToken waste — redundant calls, bloated prompts, inefficient context/output useToken waste — redundant calls, bloated prompts, inefficient context/output useself-improvementGenuine self-improvement: stella making stella more capableGenuine self-improvement: stella making stella more capable
Description
Metadata
Metadata
Assignees
Labels
P2Polish — worth doing, not urgentPolish — worth doing, not urgentarea:contextstella-context — graph, embeddings, memory, retrievalstella-context — graph, embeddings, memory, retrievalarea:pipelinestella-pipeline — triage→…→judge orchestrationstella-pipeline — triage→…→judge orchestrationfeatureNew capability or improvementNew capability or improvementgoal:enginePillar 4 — engine tuned and optimized: budget, wall-clock, resolve ratePillar 4 — engine tuned and optimized: budget, wall-clock, resolve ratepain:token-efficiencyToken waste — redundant calls, bloated prompts, inefficient context/output useToken waste — redundant calls, bloated prompts, inefficient context/output useself-improvementGenuine self-improvement: stella making stella more capableGenuine self-improvement: stella making stella more capable
Problem
The pre-plan research stage (#1778, PR #1953) fans out read-only sub-agents whose answers are bounded and rendered into the planner prompt (
crates/stella-pipeline/src/research.rs,pipeline/research_stage.rs) — then dropped with the turn. What survives is telemetry only:SubAgentbrackets andModelCallRole::Researchcalls in the session journal (store.db). Nothing reachescontext.db.A research child that just answered "where is X wired and what constrains it" produced exactly the kind of workspace fact the recall stage exists to serve. Today that knowledge is spent tokens with a one-turn lifetime; the next session re-derives it or plans without it.
What to build
Feed accepted research findings into the context plane as recallable chunks — most likely the episodic-memory ingest side of
stella-context(the same substrate reflection mining writes to), tagged with provenance (research, the question asked, the goal that prompted it, workspace + commit identity) so recall fusion can rank and expire them honestly.Design questions to settle before building:
observeddiscipline as other context records (see the staleness-guard conventions instella-context).Verify
A pipeline run whose research stage produced findings leaves chunks queryable from
context.dbwith research provenance; a later session's recall for a related goal surfaces them; a finding for a file since modified is marked stale rather than served fresh.Constraints
crates/stella-pipeline/src/pipeline.rsandpipeline/tests.rsare god files closed to growth — wiring lands in siblings.context.dbis local.Refs #1778, PR #1953.