Lap Tran
35e449b03e
Completes all 3 model-divergence enforcement subtasks:
- scripts/detect_models.py: probes opencode.json + localhost endpoints,
builds models.json with --json/--write/--force
- scripts/status.py: CONFLICT_MATRIX, --model flag, --transition --model,
--claim --model, --audit Category 6, model-divergence brake gate in
--check-gate, helpers for manifest loading and conflict checking
- scripts/loop-runner.py: _role_model() helper + {model} passed via extras
dict to _invoke_harness for implement, verify, orchestrate roles
- tests/test_model_divergence.py: 33 tests covering all enforcement layers
- Single-LLM mode: record model advisory, no conflict check
- Multi-LLM mode (2+ models): conflict matrix enforced at transition, claim,
and loop brake gate
- Project-level models.json preferred over global ~/.automaton/models.json
90 lines
3.8 KiB
Markdown
90 lines
3.8 KiB
Markdown
# Framework Configuration
|
|
|
|
This file contains global framework settings that apply across all projects.
|
|
|
|
## VRAM Configuration
|
|
|
|
Settings for task decomposition based on available VRAM.
|
|
|
|
- **Auto-detect**: Yes # Detect GPU VRAM, RAM, and model context window automatically
|
|
- **Target context**: 16k tokens # Override auto-detect if needed
|
|
- **Headroom**: 25% # Leave headroom for code, context, and reasoning
|
|
- **Max peak context per sub-task**: 12k tokens # Max context for any single sub-task
|
|
|
|
### Auto-detection
|
|
|
|
When `Auto-detect: Yes`, the framework probes your system to detect:
|
|
- GPU VRAM (via `nvidia-smi` or `lspci`)
|
|
- System RAM (via `/proc/meminfo` or `sysctl`)
|
|
- Model context window (via API config or model name lookup)
|
|
- Framework overhead (by reading all loaded prompt files)
|
|
|
|
To disable auto-detection and use manual values:
|
|
|
|
```
|
|
## VRAM Configuration
|
|
- **Auto-detect**: No
|
|
- **Target context**: 8k
|
|
- **Headroom**: 30%
|
|
- **Max peak context per sub-task**: 5.6k
|
|
```
|
|
|
|
## Model Configuration
|
|
|
|
Settings for the LLM model being used.
|
|
|
|
- **Model**: omlx/Ornith-1.0-35B-4bit-mlx # Local LLM (opencode provider); used as the Implement role
|
|
- **Override context window**: 32768 # Matches opencode.json limit.context for ornith
|
|
|
|
### Auto-detection
|
|
|
|
When `Model: auto`, the framework detects the model name from:
|
|
1. `.agent.md` in the project (if specified there)
|
|
2. API config files (`.env`, `config.yaml`, `config.json`, etc.)
|
|
3. Model name lookup by name (e.g., gpt-4o → 128k, claude-3-5-sonnet → 200k)
|
|
|
|
To disable auto-detection and use manual values:
|
|
|
|
```
|
|
## Model Configuration
|
|
- **Model**: gpt-4o
|
|
- **Override context window**: 128k
|
|
```
|
|
|
|
## Available Models
|
|
|
|
Models available for model-divergence enforcement. This file is managed by `scripts/detect_models.py`. In single-LLM mode (0-1 models), no hard blocks are enforced. In multi-LLM mode (2+ models), the conflict matrix enforces role-model separation.
|
|
|
|
- **Default**: omlx/Ornith-1.0-35B-4bit-mlx # Used when no role-specific binding is set
|
|
- **Advised**: true # Recommend a second model in single-LLM mode
|
|
|
|
No additional models are configured in the manifest. To add models:
|
|
1. Run `python3 ~/.automaton/scripts/detect_models.py --write` to auto-detect from opencode.json and localhost endpoints.
|
|
2. Or manually create `~/.automaton/models.json` (see `design/framework/technical.md` §2 for schema).
|
|
|
|
## System Requirements
|
|
|
|
Requirements for the environment the framework runs in.
|
|
|
|
- **nvidia-smi**: Required if NVIDIA GPU (for VRAM detection)
|
|
- **lspci**: Fallback if NVIDIA GPU not available (for AMD GPU VRAM detection)
|
|
- **/proc/meminfo**: Required for RAM detection (Linux)
|
|
- **sysctl**: Fallback for RAM detection (macOS)
|
|
|
|
## Loop Role Models
|
|
|
|
Loop ticks run three session roles. Roles are *sessions*, not models — a single model can fill multiple roles. Configure each loop's role-to-prompt binding in its `loop.json`; this section documents the framework's expectations only.
|
|
|
|
- **Implement:** — produces the artifact for this tick. Bound to `prompts/loop-implement.md` by default.
|
|
- **Verify:** — grades the artifact and emits the JSON verdict `{pass, score, reasons, next_hint}`. Bound to `prompts/loop-verifier.md`. The framework never inspects this role's model (D8); only its session.
|
|
- **Orchestrate:** — applies the verdict, calls exactly one `status.py` operation per tick, enforces brakes. Bound to `prompts/loop-orchestrate.md`.
|
|
|
|
Conflict-of-interest rule (D12): `Verify:` and `Implement:` must never be the same *session*. When two distinct sessions are infeasible (single-session harness), the runner falls back to session-only divergence — still safe.
|
|
|
|
Role context tiers are set per-loop in `loop.json`, not globally. The 16k floor (D13) applies regardless of tier.
|
|
|
|
## Framework Version
|
|
|
|
- **Version**: 2.0
|
|
- **State enforcement**: enabled (`.state` file + `status.py`)
|