@nanobpm/nano-coder 0.40.0 → 0.42.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (2) hide show
  1. package/README.md +43 -0
  2. package/package.json +5 -5
package/README.md CHANGED
@@ -54,6 +54,16 @@ line shows how many are waiting. Edit the queue at any time
54
54
  drops one (or several, `/queue remove N M`), `/queue edit N new text` rewrites one, and
55
55
  `/queue clear` empties the queue. A queued message is never injected while the agent is
56
56
  asking for input (the `question` tool's picker owns the terminal until you answer).
57
+ Slash commands work mid-turn too, where it is safe:
58
+ - **Right away:** read-only ones (`/help`, `/tools`, `/skills`, `/providers`, `/session`, `/plan`,
59
+ `/context`, `/trajectory` (`--json`/`--markdown`), and `/mode` or `/verbosity` without an
60
+ argument) show the current state. `/plan` shows the plan as the agent updates it.
61
+ - **From the agent's next step:** `/mode NAME` applies when the agent takes its next step.
62
+ - **Right away (for the rest of the turn):** `/verbosity LEVEL` changes the output level
63
+ immediately, so it affects the remaining events streamed by the current turn.
64
+ - **After the turn:** commands that change the conversation or take over the keyboard
65
+ (`/compact`, `/restart`, `/settings`, `/model`, `/exit`) wait for the turn to finish.
66
+
57
67
  **Esc Esc** (twice within a second) or **Ctrl-C** cancels
58
68
  the running turn, killing any running bash command; a second Ctrl-C at the prompt exits.
59
69
  At the idle prompt, **Esc Esc** instead clears the input, so a half-typed or pasted prompt
@@ -679,6 +689,39 @@ reaches the wire. The status bar shows
679
689
  `/settings` sets it for the current model, its provider or all models (picking from the
680
690
  levels the model supports), and the session log records it for each reply.
681
691
 
692
+ ### Vision
693
+
694
+ `read_file` returns images (PNG, JPEG, GIF and WebP, recognised by magic bytes) to
695
+ models that can view them: the result is a short text part (`image/png, 1600×900, 131 KB`)
696
+ plus an image attachment the model sees directly. Images over the provider's limits are
697
+ downscaled first (longest side 1568 px, under the model's byte cap), and re-encoded to
698
+ JPEG/PNG when the model doesn't accept the source type. Other binary files still return
699
+ the "looks like a binary file" error, which notes whether the current model supports
700
+ images.
701
+
702
+ Whether a model can see images comes from, in order: a `vision = true|false` override
703
+ (global, `[providers.<name>]`, or `[providers.<name>.models."<model>"]`, like `thinking`);
704
+ what the endpoint reports (GitHub Copilot `/models` `capabilities.supports.vision` and its
705
+ `limits.vision`, Ollama `/api/show`'s `vision` capability, llama.cpp `/props`
706
+ `modalities.vision`); and a built-in assumption for current Anthropic and OpenAI model
707
+ families. A model that can't see images gets the text error with a hint to switch models
708
+ or set `vision = true`:
709
+
710
+ ```toml
711
+ vision = true # all models
712
+
713
+ [providers.ollama.models."my-clip-model"]
714
+ vision = true # this model only
715
+ ```
716
+
717
+ A request carries only the newest few images the model allows (GitHub Copilot's
718
+ `max_prompt_images`, default 1); older ones become `[image omitted: path (sent earlier)]`.
719
+ Each image counts toward the context estimate at a fixed cost from its size, so the status
720
+ bar and auto-compaction stay honest, and compaction replaces images with the same
721
+ placeholder. Images are stored once per session under `<session>.attachments/` (named by
722
+ content hash), so session logs stay small and `--resume` still works; a missing file on
723
+ resume is sent as a text placeholder.
724
+
682
725
  Other per-provider fields: `replay_reasoning`, `max_tokens_param` (`max_tokens`, or `max_completion_tokens`
683
726
  which is the `openai` default), `retry_initial_backoff_ms`, `retry_max_backoff_ms` and
684
727
  `retryable_statuses`.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@nanobpm/nano-coder",
3
- "version": "0.40.0",
3
+ "version": "0.42.0",
4
4
  "description": "A 6MB coding agent. Run a fleet on your laptop.",
5
5
  "license": "Apache-2.0",
6
6
  "repository": {
@@ -31,9 +31,9 @@
31
31
  "node": ">=18"
32
32
  },
33
33
  "optionalDependencies": {
34
- "@nanobpm/nano-coder-darwin-arm64": "0.40.0",
35
- "@nanobpm/nano-coder-darwin-x64": "0.40.0",
36
- "@nanobpm/nano-coder-linux-arm64": "0.40.0",
37
- "@nanobpm/nano-coder-linux-x64": "0.40.0"
34
+ "@nanobpm/nano-coder-darwin-arm64": "0.42.0",
35
+ "@nanobpm/nano-coder-darwin-x64": "0.42.0",
36
+ "@nanobpm/nano-coder-linux-arm64": "0.42.0",
37
+ "@nanobpm/nano-coder-linux-x64": "0.42.0"
38
38
  }
39
39
  }