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).
- Updated all 23 skill org files to use absolute path
- Tangle now outputs directly to ~/.local/share/opencortex/skills/
- Removed env var (org-babel doesn't expand it)
- Updated all 22 skill org files to use $OC_DATA_DIR/skills/ paths
- Removed manually created .lisp file (tangling now targets XDG)
- Files will now tangle to ~/.local/share/opencortex/skills/
Self-edit: 5 new tests (apply success/not-found/file-not-found, parse-location x2)
Config-manager: 4 new tests (get-oc-config-dir, save-providers, configure-provider)
Gateway-manager: 2 new tests (multiple-platforms, registration)
Tier 1 Chaos: Verified org files pass structural balance
Note: Some tests have issues - config tests use functions not exported, one self-edit test has search function issue. Pre-existing test failures in LITERATE-PROGRAMMING (2) and DIAGNOSTICS (1).
- Standardized def-cognitive-tool to 5-argument signature.
- Consolidated *cognitive-tools* as a hash table in package.lisp.
- Removed skills from opencortex.asd to enforce dynamic Skill Engine loading.
- Added missing (in-package :opencortex) to various skill files.
- Fixed let/let* sequential binding issues in emacs-edit and self-edit.
- Updated opencortex.sh to initialize skills before running doctor.
- Fixed uiop:user-homedir-pathname usage in config-manager.
- 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.
- Hardened actuators: Fixed path-traversal vulnerabilities in file I/O tools and blocked .org files from regex replacements to force AST usage. Enforced Merkle snapshots on AST edits.
- Implemented Reflection Loops: Injected rejection traces from deterministic gates back into the LLM context to enable autonomous self-correction.
- Finalized tool permission tiers (ask/allow/deny) with proper LLM prompt filtering.
- Deleted docs/rca/ as they were early v0.1.0 development artifacts.
- Deleted minimal.asd.
- Deleted scripts/onboard-baremetal.sh (subsumed by opencortex.sh).
- Moved scripts/browser-bridge.py to skills/assets/ for upcoming Web Research skill.
- Removed scripts/ directory.
- 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.)
- Removed *tangle-targets* and check-tangle-sync from engineering-standards.org
- Fixed LP org: added (in-package), fixed block balance, fixed return-from
- Fixed literate-check-block-balance to return (values nil reason) on imbalance
- Updated LP tests to work with *tangle-targets* override
- Regenerated all .lisp from org via emacs --batch
- Added LP gen back to opencortex.asd
Test results:
- Engineering Standards: 9/10 (90%)
- Literate Programming: 7/7 (100%)
- Restored engineering-standards.org from 86eeaab (was missing blocks)
- Regenerated .lisp from org (proper literate workflow)
- Both org and .lisp now have correct paren balance
- System loads, tests pass 9/10 (1 expected failure due to dirty git)
Lesson: NEVER edit .lisp directly. Always fix org and regenerate.
Engineering Standards now focuses on lifecycle phases (0, A, B, D, E).
Literate Programming now owns LP structural invariants including tangle-sync.
Changes:
- Removed check-tangle-sync and *enforcement-rules* from org-skill-engineering-standards.org
- Added check-tangle-sync, *tangle-targets*, *lp-project-root* to org-skill-literate-programming.org
- Updated LP skill to check tangle-sync on file modification actions
- Added literate-programming-tests.lisp with tangle-sync and block-balance tests
- Removed tangle-sync tests from engineering-standards-tests.lisp
