Replace Ollama-specific backend with unified org-skill-unified-llm-backend
that speaks OpenAI API. Works with:
- Local: Ollama (default), vLLM, LM Studio, llama.cpp
- Cloud: OpenRouter, OpenAI, Anthropic, Groq, Gemini
Providers auto-registered from env vars. No separate skills per provider.
Cascade order configured via PROVIDER_CASCADE env var.
Also fix .env loading path in loop (was .local/share, now .config matches wizard).
- Finalized Reflection Loop: Injected deterministic rejection traces back into LLM prompts.
- Hardened Actuators: Added path-traversal guards and enforced Merkle snapshots on AST edits.
- Refactored Lisp Utils: Merged validator/repair into a unified utility skill with whitelist Ast-walking.
- Fixed Build: Resolved all 30+ syntax, scoping, and package visibility errors.
- Verified: Full pass (100%) on all 5 core test suites.
- Updated all 150+ :tangle headers across harness/ and skills/ to use elisp (expand-file-name) to target INSTALL_DIR dynamically.
- Cleaned up environment/ directory depth by moving memory-image.lisp to state/.
- Moved test scripts to tests/ and deleted redundant chat scripts.
- Fix critical paren balance issues across harness/skills.org, act.org,
loop.org, memory.org, and skills/self-edit|emacs-edit.org
- Add :reload-skill cognitive tool for hot-reloading without restart
- Add :generate-embeddings tool and self-edit hot-reload infrastructure
- Wire all new skills (self-edit, emacs-edit, lisp-utils) into main ASDF
- Regenerate all .lisp tangled files via emacs --batch org-babel-tangle
- Add :opencortex/tests ASDF system with 14 test suites
- Fix test files to compile cleanly (self-edit-tests symbol vis, etc.)
- Add test suites inline in harness/communication.org
- Add test suites inline in harness/perceive.org, reason.org, act.org
- Add test suites inline in harness/skills.org, memory.org, loop.org
- Tests now live alongside code, not in separate .lisp files
- Each test block has prose explaining its purpose
Test results: 84/84 pass (100%)
- loop.org: Explain metabolic pipeline, error recovery, heartbeat
- reason.org: Document dual-engine architecture (probabilistic + deterministic)
- act.org: Detail actuation, actuator pattern, feedback loop
- perceive.org: Explain signal normalization, async vs sync processing
- manifest.org: Add ASDF system design, boundary contract
Each function now has its own code block with detailed docstrings explaining:
- What the function does
- Why it was designed that way
- How it fits into the larger system
- Parameter descriptions
- Return values
- Add save-memory-to-disk and load-memory-from-disk to memory.lisp
- Integrate auto-save into heartbeat (every N intervals)
- Load memory on daemon startup, save on graceful shutdown/SIGINT
- Add exports to package.lisp
NOTE: Hash table serialization requires object walker for complex structures.
Current implementation fails on load due to unreadable objects.
