Files

T

CI / build (push) Has been cancelled

Details

Flatten model-divergence subtasks into 3 independent tasks

Replaced parent+subtask structure with 3 standalone tasks:
- mde-manifest-detection (foundational: models.json schema, detect_models.py)
- mde-interactive-enforcement (conflict matrix, --model args, audit, badges)
- mde-loop-enforcement (loop.json per-role model, {model} substitution, check-gate)

Parent archived to tasks/complete/ for history. Each task has its own
SPEC.md and is at research:awaiting_approval.

Rationale: subtasks are independently actionable with their own lifecycle.
Parent container added complexity without benefit.

2026-06-25 07:25:10 -04:00

2.0 KiB

Raw Blame History

SPEC — mde-manifest-detection

Problem

The framework has no model manifest system. Model-divergence enforcement (conflict matrix, per-role model binding, loop enforcement) requires knowing which models are available and whether the system is in single-LLM or multi-LLM mode.

Design

Full design in design/framework/functional.md §4 and design/framework/technical.md §§2-3.

Key decisions

Manifest: models.json with {default, advised, models:[{name, provider, context_window, location}]}.
Mode detection: 0-1 models → single-LLM (advisory once, then silent). 2+ → multi-LLM (hard block). Missing file → single-LLM (backward compatible).
Detection: scripts/detect_models.py probes opencode.json providers + localhost endpoints (8080/11434/1234/8000).

Scope

scripts/detect_models.py — probe opencode.json providers + localhost endpoints, emit candidate manifest as JSON to stdout
scripts/status.py — add _load_models_manifest() and _get_mode() helpers
scripts/install.sh, update.sh, upgrade.sh — call detect_models after vram_detect
config.md — ## Available Models section template
prompts/onboarding.md — Step 2d model config check
tests/test_model_divergence.py — manifest loading, single vs multi mode, missing file

Out of scope

Conflict matrix enforcement (subtask mde-interactive-enforcement)
Loop model binding (subtask mde-loop-enforcement)
Rule agents (FW-2, FW-3) — separate backlog items
Agent tab redesign (FW-1)

Success criteria

models.json missing → _get_mode() returns "single", all model commands are no-ops
models.json with 0-1 models → "single" mode
models.json with 2+ models → "multi" mode
detect_models.py probes localhost and prints candidate JSON
pytest tests/ -q green

References

design/framework/functional.md §4 (Model-Divergence Enforcement)
design/framework/technical.md §§2-3 (manifest schema, state schemas)
design/framework/README.md (locked decisions F4, F5)

2.0 KiB Raw Blame History