8000 chore: migrate io-auto-mode classifier defaults to gemini-3.1-flash-lite by simon-inkie · Pull Request #1 · simon-inkie/inkie-auto-mode · GitHub
[go: up one dir, main page]

Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 11 additions & 10 deletions INSTALL.md
Original file line number Diff line number Diff line change
Expand Up @@ -18,8 +18,9 @@ Pick whichever runtime you use; they all share the same `core/` classifier.
## Requirements

- Node.js 22+
- A Google Gemini API key (defaults use `gemini-2.5-flash` for all stages —
fast, cheap, and accurate enough for classification). You can swap providers
- A Google Gemini API key (defaults use `gemini-3.1-flash-lite` for both Stage 1
and the Stage 2 thinking classifier: fast, cheap, and accurate enough for
classification). You can swap providers
via config; see Configuration Reference below.
- For the OpenClaw adapter: OpenClaw **2026.4.2+**.

Expand Down Expand Up @@ -429,10 +430,10 @@ Add to `plugins.load.paths` so OpenClaw discovers it on startup:
"enabled": true,
"config": {
"mode": "classify",
"stage1Model": "google/gemini-2.5-flash",
"stage1Fallback": "google/gemini-2.5-flash",
"stage2Model": "google/gemini-2.5-flash",
"stage2Fallback": "google/gemini-2.5-flash"
"stage1Model": "google/gemini-3.1-flash-lite",
"stage1Fallback": "google/gemini-3.1-flash-lite",
"stage2Model": "google/gemini-3.1-flash-lite",
"stage2Fallback": "google/gemini-3.1-flash-lite"
}
}
}
Expand Down Expand Up @@ -479,10 +480,10 @@ All options live under `plugins.entries.io-auto-mode.config`:
| Option | Default | Description |
|--------|---------|-------------|
| `mode` | `classify` | `classify` (normal), `yolo` (allow all), `strict` (block unless on allowlist) |
| `stage1Model` | `google/gemini-2.5-flash` | Fast LLM for Stage 1 classification |
| `stage1Fallback` | `google/gemini-2.5-flash` | Fallback if Stage 1 model unavailable |
| `stage2Model` | `google/gemini-2.5-flash` | Thinking LLM for Stage 2 (escalated blocks) |
| `stage2Fallback` | `google/gemini-2.5-flash` | Fallback if Stage 2 model unavailable |
| `stage1Model` | `google/gemini-3.1-flash-lite` | Fast LLM for Stage 1 classification |
| `stage1Fallback` | `google/gemini-3.1-flash-lite` | Fallback if Stage 1 model unavailable |
| `stage2Model` | `google/gemini-3.1-flash-lite` | Thinking LLM for Stage 2 (escalated blocks) |
| `stage2Fallback` | `google/gemini-3.1-flash-lite` | Fallback if Stage 2 model unavailable |
| `userAllowPatterns` | `[]` | Additional regex patterns to always allow |
| `userBlockPatterns` | `[]` | Additional regex patterns to always block |

Expand Down
8 changes: 4 additions & 4 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -278,10 +278,10 @@ All options under `plugins.entries.io-auto-mode.config`:
| Option | Default | Description |
|---|---|---|
| `mode` | `classify` | `classify` / `yolo` / `strict` |
| `stage1Model` | `google/gemini-2.5-flash` | Fast LLM for Stage 1 |
| `stage1Fallback` | `google/gemini-2.5-flash` | Stage 1 fallback |
| `stage2Model` | `google/gemini-2.5-flash` | Thinking LLM for Stage 2 |
| `stage2Fallback` | `google/gemini-2.5-flash` | Stage 2 fallback |
| `stage1Model` | `google/gemini-3.1-flash-lite` | Fast LLM for Stage 1 |
| `stage1Fallback` | `google/gemini-3.1-flash-lite` | Stage 1 fallback |
| `stage2Model` | `google/gemini-3.1-flash-lite` | Thinking LLM for Stage 2 |
| `stage2Fallback` | `google/gemini-3.1-flash-lite` | Stage 2 fallback |
| `userAllowPatterns` | `[]` | Extra regex patterns always allowed |
| `userBlockPatterns` | `[]` | Extra regex patterns always blocked |

Expand Down
8 changes: 4 additions & 4 deletions core/types.ts
Original file line number Diff line number Diff line change
Expand Up @@ -47,10 +47,10 @@ export interface ClassifierConfig {

/** Default configuration */
export const DEFAULT_CONFIG: ClassifierConfig = {
stage1Model: 'google/gemini-2.5-flash',
stage1Fallback: 'google/gemini-2.5-flash',
stage2Model: 'google/gemini-2.5-flash',
stage2Fallback: 'google/gemini-2.5-flash',
stage1Model: 'google/gemini-3.1-flash-lite',
stage1Fallback: 'google/gemini-3.1-flash-lite',
stage2Model: 'google/gemini-3.1-flash-lite',
stage2Fallback: 'google/gemini-3.1-flash-lite',
mode: 'classify',
nonMainMode: 'block',
userAllowPatterns: [],
Expand Down
Loading
0