Back to corpus
proposalexperiment writeup candidatescore 22

KARL Integration — Evolution³ / Stage 1: PATH E

Path E adapts KARL's synthetic self-play pipeline to our living codebase. Instead of mining static enterprise documents, our question generator reads our own code, memory files, hooks, flows, and configs to produce domain-specific questions. A solver agent then attempts to answer each question using our actual tool stack (Read, Grep, Bash, RAG++, GK). Every attempt is recorded as a trajectory. Trajectories are quality-filtered, then used to either improve SKILL.md content (near-term, zero training cost) or train a

Open full HTMLDownload HTMLCorpus JSONPaper schema

Full HTML reader

Read the full artifact

Open in new tab

Extracted abstract or opening context

# KARL Integration — Evolution³ / Stage 1: PATH E **Run:** karl-trajectory-intelligence **Path:** E — Synthetic Self-Play: Generate Our Own Training Data from the Codebase **Stage:** 1 of 4 (Explore + Design) **Generated:** 2026-03-10 **Method:** Evolution³ — four-stage recursive evoflow (research-grounded) **Run Directory:** Desktop/evo-cube-output/karl-trajectory-intelligence/ Path E adapts KARL's synthetic self-play pipeline to our living codebase. Instead of mining static enterprise documents, our question generator reads our own code, memory files, hooks, flows, and configs to produce domain-specific questions. A solver agent then attempts to answer each question using our actual tool stack (Read, Grep, Bash, RAG++, GK). Every attempt is recorded as a trajectory. Trajectories are quality-filtered, then used to either improve SKILL.md content (near-term, zero training cost) or train a LoRA adapter on Mac5 (medium-term, higher yield). The net result is a self-improving knowledge base that gets better the more it is used, without requiring any manual documentation work. **Why this is the right path over alternatives:** - Path A (add OAPL directly) requires distributed GPU training we do not have. - Path B (plug in KARL's external model weights) gives us someone else's knowledge, not ours. - Path C (add reward signals to existing Cortex) improves routing but not knowledge depth. - Path D (use RAG++ as the KARL search tool) improves retrieval but not the agent's procedures. - **Path E** generates our own proprietary training signal from our own codebase — a moat no external model can replicate. The question generator mines five tiers of source material, ordered by information density: | Tier | Source | Volume | Why | |------|--------|--------|-----| | T1 | `[home-path]` topic files | 29 files, ~3,500 lines | Curated operational knowledge with hard-won gotchas | | T2 | `[home-path]` SKILL.md files (Gen 2 only) | 12 active, ~800 lines | Procedure documents with trigger patterns and workflows | | T3 | `flows/feed-hub/*.py` — Prefect flow source | 106 files | Actual implementation truth for flow/task questions | | T4 | `[home-path]` — all hook source files | 34 hooks, 29 scripts | Ground truth for hook behavior questions | | T5 | `[home-path]` (filtered) | ~400 real human prompts after SKIP_PATTERNS | Reveals what humans actually need to know |

Promotion decision

What has to happen next

Attach run IDs, datasets, metrics, and reproduction commands.

Why this is not always a full paper yet

Corpus pages are public-safe readers for discovered workspace artifacts. They are not automatically final papers. A corpus item becomes a polished paper only after the editable source, evidence checkpoints, references, figures, render path, and release status are attached through the paper schema.