Skip to content

armgit / ruflo

Public Imported from GitHub
0

🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated

darwin: scaffold autonomous capability-evolution loop (PLAN + log)

ruv1b1e20ec38d2
diff --git a/docs/darwin/PLAN.md b/docs/darwin/PLAN.md
new file mode 100644
index 0000000000..2c07e264a1
--- /dev/null
+++ b/docs/darwin/PLAN.md
@@ -0,0 +1,52 @@
+# Darwin capability evolution β€” plan
+
+Branch: `darwin/capability-evolution-2026-06-26`
+Started: 2026-06-26
+
+## Goal
+Drive ruflo capabilities toward SOTA across the dimensions we already
+benchmark, using a `/loop 5m` autonomous loop. Each tick spawns one
+`claude -p` (headless, Read/Edit/Bash only, --max-budget-usd capped) to
+do a single optimization cycle, so this conversation stays focused on
+orchestration and the per-tick spend is bounded.
+
+## Per-tick contract
+A single tick = one `claude -p` invocation that does end-to-end:
+
+ 1. Read `docs/darwin/log.jsonl` β€” last N entries, find current
+ champion scores per dimension.
+ 2. Pick the worst-relative-to-SOTA dimension. SOTA baselines:
+ BEIR NFCorpus β€” nDCG@10 β‰₯ 0.36 (state-of-the-art hybrid)
+ BEIR ArguAna β€” nDCG@10 β‰₯ 0.55
+ BEIR SciFact β€” nDCG@10 β‰₯ 0.74
+ BEIR TREC-COVID β€” nDCG@10 β‰₯ 0.78
+ GAIA L1 β€” exact-match β‰₯ 0.62 (LangGraph reference)
+ ADR coverage β€” adr-index storage success β‰₯ 0.99
+ 3. Propose ONE targeted change (parameter tune, prompt rewrite, dep
+ bump, algorithm swap). Keep it small enough that a benchmark
+ subset can score it in ≀4 minutes.
+ 4. Apply, run the relevant benchmark/audit:
+ BEIR β†’ `node scripts/bench-beir.mjs <dataset> --top-k 10`
+ ADR β†’ `node plugins/ruflo-adr/scripts/import.mjs --dry-run`
+ OIA β†’ `npx ruflo metaharness oia-audit --format json`
+ 5. Compare delta to prior champion for that dimension:
+ Ξ” > 0 β†’ commit, update champion, log success
+ Ξ” ≀ 0 β†’ revert, log noImprovement
+ benchmark error β†’ revert, log error
+ 6. Append one JSONL line to `docs/darwin/log.jsonl` with:
+ { iter, ts, dimension, change, deltaScore, action, commit }
+
+## Termination
+- 3 consecutive iterations without Ξ” > 0 across any dimension β†’ stop
+- Or explicit user stop
+
+## Spend cap per tick
+`claude -p --max-budget-usd 0.50 --model haiku` for routine ticks.
+Escalate to sonnet only when haiku reports "task too complex" 3x in
+a row.
+
+## SOTA-proof
+A dimension is "proven SOTA" when:
+ - It exceeds the baseline above by β‰₯1%
+ - The benchmark run is reproducible (3 consecutive runs within 1Οƒ)
+ - The git commit is signed and witnessed
diff --git a/docs/darwin/log.jsonl b/docs/darwin/log.jsonl
new file mode 100644
index 0000000000..e69de29bb2