Rename the athena skill to compile — a descriptive name that communicates
what it actually does. Extend with Karpathy's LLM Knowledge Bases architecture:
- New compile phase: LLM reads raw .agents/ artifacts and produces interlinked
markdown wiki at .agents/compiled/ with index.md, log.md, backlinks
- New lint phase: detects contradictions, orphan pages, stale claims, coverage gaps
- Pluggable compute backend via AGENTOPS_COMPILE_RUNTIME (ollama|claude|openai)
- Hash-based incremental compilation (only recompile changed sources)
- compile.sh engine with security fixes (no shell injection, path traversal guard)
- Full rename across 93 files: skills, docs, CLI, tests, hooks, scripts, codex
Existing Mine → Grow → Defrag cycle preserved as phases within the larger
compile workflow. No vector DB — compiled wiki IS the retrieval layer.
Map runtime mechanics in how-it-works.md and knowledge-flywheel.md to the
canonical three gaps (judgment validation, durable learning, loop closure).
Add gap annotations to hook tables, flywheel stages, and compounding effect.
Replace generic language with AgentOps-specific terminology.
- Update flywheel skill SKILL.md to remove --golden flag references
- Update README.md and docs/knowledge-flywheel.md
- Rename test to TestFlywheelStatus_GoldenSignalsAlwaysShown and verify
golden signals appear without the flag
Four derived health indicators that distinguish knowledge compounding
from noise accumulation:
1. Velocity Trend — is σρ-δ increasing over time?
2. Citation Pipeline — are citations delivering value?
3. Research Closure — is research being mined into learnings?
4. Reuse Concentration — is the whole pool active or just a few items?
Accessible via `ao flywheel status --golden` (JSON/YAML/table).
Also adds forge→pool bridge (forge_pending.go) that writes pending
learnings as markdown for close-loop ingestion, and session-end
baseline capture hook for velocity trend tracking.
Security fixes from council review: path traversal sanitization,
frontmatter injection prevention, YAML encoder resource leak.
Boden Fuller