oioxo-mcp 0.3.1 → 0.4.1
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.
- package/README.md +6 -6
- package/bundle/cli.js +140 -91
- package/package.json +3 -2
package/README.md
CHANGED
|
@@ -4,7 +4,7 @@
|
|
|
4
4
|
|
|
5
5
|
OIOXO does two things from one CLI:
|
|
6
6
|
1. **Context engine** — sits between your codebase and your AI agents (GitHub Copilot, Claude Code, Cursor, Windsurf, Gemini CLI, Codex) and hands them the *minimal relevant slice* of your project instead of letting them read whole files.
|
|
7
|
-
2. **Coding agent** — `oioxo code "<task>"` plans, edits, and verifies in your repo, using your own key **or a
|
|
7
|
+
2. **Coding agent** — `oioxo code "<task>"` plans, edits, and verifies in your repo, using your own key **or a local Ollama model** — nothing leaves your machine. See [Code in your terminal](#code-in-your-terminal--oioxo-code).
|
|
8
8
|
|
|
9
9
|
**Measured on a real production codebase: 90–92% fewer context tokens per question.** A question that costs an agent ~50,000 tokens of file reading comes back as a ~5,000-token capsule — same answer, a fraction of the cost.
|
|
10
10
|
|
|
@@ -30,15 +30,15 @@ OIOXO isn't only a context engine for *other* agents — it's a coding agent in
|
|
|
30
30
|
OPENAI_API_KEY=… oioxo code "add a --json flag to the status command"
|
|
31
31
|
ANTHROPIC_API_KEY=… oioxo code "fix the failing test" --verify "npm test"
|
|
32
32
|
|
|
33
|
-
#
|
|
34
|
-
oioxo code "add input validation to the signup form"
|
|
33
|
+
# local Ollama — private, zero API tokens, nothing leaves your machine
|
|
34
|
+
OIOXO_PROVIDER=ollama OIOXO_MODEL=qwen2.5-coder oioxo code "add input validation to the signup form"
|
|
35
35
|
```
|
|
36
36
|
|
|
37
37
|
- **Verified, not vibes** — every change is run through your tests/typecheck on a shadow copy; a red result is fed back and repaired. Your real files are only touched after you approve the diff.
|
|
38
38
|
- **Bring any model** — OpenAI, Anthropic, Groq, Mistral, Together, local Ollama, or any OpenAI-compatible base (`OIOXO_PROVIDER` / `OIOXO_BASE_URL` / `OIOXO_MODEL` / `OIOXO_API_KEY`).
|
|
39
|
-
- **Or
|
|
39
|
+
- **Or fully local, no key** — point it at [Ollama](https://ollama.com) (`OIOXO_PROVIDER=ollama OIOXO_MODEL=qwen2.5-coder`) and nothing leaves your machine. A built-in `--local` on-device coder (open model via llama.cpp, auto-downloaded once to `~/.oioxo/models`) is **coming next**.
|
|
40
40
|
|
|
41
|
-
Flags: `--verify "<cmd>"` (override the test command) · `--yes` (apply without the prompt) · `--local`
|
|
41
|
+
Flags: `--verify "<cmd>"` (override the test command) · `--yes` (apply without the prompt) · `--max-iters N`. *(`--local` is reserved for the built-in on-device coder, coming next — use `OIOXO_PROVIDER=ollama` for local today.)*
|
|
42
42
|
|
|
43
43
|
## Why developers use it
|
|
44
44
|
|
|
@@ -66,7 +66,7 @@ Flags: `--verify "<cmd>"` (override the test command) · `--yes` (apply without
|
|
|
66
66
|
|
|
67
67
|
```
|
|
68
68
|
oioxo-mcp code "<task>" code in this repo: capsule-grounded, verified, then asks
|
|
69
|
-
before writing (--verify "<cmd>" · --yes · --
|
|
69
|
+
before writing (--verify "<cmd>" · --yes · --max-iters)
|
|
70
70
|
oioxo-mcp login connect your OIOXO account
|
|
71
71
|
oioxo-mcp init [--all] write agent MCP configs + instruction files
|
|
72
72
|
oioxo-mcp serve run the MCP server (agent configs call this)
|