vtx-coding-agent 0.2.4__tar.gz → 0.2.5__tar.gz

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/AGENTS.md

@@ -26,21 +26,13 @@
 - If the user asks you to push code, run these first before doing so: `uv run ruff format .`, `uv run ruff check .`, `uvx ty check  .` and `uv run python -m pytest` in parallel (same tool call)
 - Only if these all pass without issues should you push otherwise report the warnings/errors back to user and ask for next steps
 
-## CodeBase Search
+## Codebase Search
 
-Search the codebase using semantic + BM25 hybrid retrieval. PREFER this over grep/rg/Glob for finding code.
+Use vortexa (not grep/rg/file reads) to search code or understand a repo. It
+indexes the current directory (or pass --root <dir>).
 
-Use this tool when you need to:
-- Find where a function, class, or pattern is implemented
-- Understand how a feature or concept works across the codebase
-- Locate code by describing what it does (not just exact strings)
-- Find examples of a pattern or API usage
-- Explore unfamiliar parts of the codebase
-- Identify relevant code for a task without knowing exact names or locations
-- Preferable than grep/rg/Glob for code search because it can understand the meaning and intent behind the query, rather than just matching exact strings. This allows for more flexible and powerful searching, especially in large and complex codebases.
-Advantages over text search: understands synonyms, paraphrases, and intent.
-Returns file paths, line ranges, relevance scores, and matching code.
+  vortexa resolve "<query>" --plain          # default: matches + tests + callers/callees + deps
+  vortexa search "<query>" --hybrid --plain  # ranked hits + per-file graph context
+  vortexa explain "<file>:<line>|<symbol>"   # deep dive into a known location
 
-- For codebase Search use `vortexa -q "you natural query here"` to get relevant code snippets from the codebase. This will help you understand the existing code and avoid duplicating functionality.
-
-- Instead of using grep use rg (ripgrep) for faster and more efficient searching. For example, `rg "def my_function"` to find all occurrences of a function definition.
+Install: pip install vortexa  (add [full] for tree-sitter AST chunking).

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/CHANGELOG.md

@@ -5,6 +5,16 @@ All notable changes to Vtx are documented in this file. The format is based on
 adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 
+## [0.2.5] - 2026-07-12
+
+### Added
+- **Meta AI provider** — registered Meta AI as a new OpenAI-compatible provider (slug `meta`) with base URL `https://api.meta.ai/v1`, resolving `META_API_KEY` from the environment, and fetching its Llama model catalog from `/models`.
+- **AGENTS.md codebase search** — replaced the CodeBase Search section with `vortexa` CLI usage (`resolve` / `search --hybrid` / `explain`) so the agent prefers a semantic index over grep/rg for code search.
+
+### Fixed
+- **Typing** — widened `_slim_schema()` parameter and return types from `object` to `Any` in `src/vtx/tools/__init__.py` to satisfy `ty`.
+- **System-prompt migration tests** — updated stale assertions from "You are Vtx, an expert coding assistant." to "You are Vtx, an expert coding agent." in `tests/test_config_migration.py`.
+
 ## [0.2.4] - 2026-07-11
 
 ### Added
@@ -868,6 +878,8 @@ to keep the model's context window free for what matters.
   via `uv tool install vtx-coding-agent`.
 - Licensed under **Apache License 2.0**.
 
+[0.2.5]: https://github.com/OEvortex/vtx-coding-agent/compare/v0.2.4...v0.2.5
+[0.2.4]: https://github.com/OEvortex/vtx-coding-agent/compare/v0.2.3...v0.2.4
 [0.2.2]: https://github.com/OEvortex/vtx-coding-agent/compare/v0.2.1...v0.2.2
 [0.2.1]: https://github.com/OEvortex/vtx-coding-agent/compare/v0.2.0...v0.2.1
 [0.2.0]: https://github.com/OEvortex/vtx-coding-agent/compare/v0.1.9...v0.2.0

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vtx-coding-agent
-Version: 0.2.4
+Version: 0.2.5
 Summary: Minimalist coding agent harness with a Textual TUI and headless CLI. Ships two agentic backends: the native vtx event loop and the advanced vtx-claw gateway loop (install the [claw] extra for multi-channel + subagent + cron features). <1k-token system prompt, 50+ LLM providers, AGENTS.md + skills context, session tree, prompt/auto permissions.
 License-File: LICENSE
 Keywords: agent-loop,cli,coding-agent,llm,tui,vtx-claw
@@ -86,7 +86,10 @@ Description-Content-Type: text/markdown
 
 <div align="center">
 
-```text
+<table align="center">
+<tr><td align="center">
+
+```
 ██╗   ██╗████████╗██╗  ██╗
 ██║   ██║╚══██╔══╝╚██╗██╔╝
 ██║   ██║   ██║    ╚███╔╝
@@ -95,8 +98,10 @@ Description-Content-Type: text/markdown
   ╚═══╝     ╚═╝   ╚═╝  ╚═╝
 ```
 
-</div>
+</td></tr>
+</table>
 
+</div>
 
 <p align="center"><b>The minimalist, modular coding agent harness</b></p>
 <p align="center"><b>Maximum capability. Minimum overhead.</b></p>

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/README.md

@@ -1,6 +1,9 @@
 <div align="center">
 
-```text
+<table align="center">
+<tr><td align="center">
+
+```
 ██╗   ██╗████████╗██╗  ██╗
 ██║   ██║╚══██╔══╝╚██╗██╔╝
 ██║   ██║   ██║    ╚███╔╝
@@ -9,8 +12,10 @@
   ╚═══╝     ╚═╝   ╚═╝  ╚═╝
 ```
 
-</div>
+</td></tr>
+</table>
 
+</div>
 
 <p align="center"><b>The minimalist, modular coding agent harness</b></p>
 <p align="center"><b>Maximum capability. Minimum overhead.</b></p>

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/pyproject.toml

@@ -20,7 +20,7 @@ default = true
 
 [project]
 name = "vtx-coding-agent"
-version = "0.2.4"
+version = "0.2.5"
 description = "Minimalist coding agent harness with a Textual TUI and headless CLI. Ships two agentic backends: the native vtx event loop and the advanced vtx-claw gateway loop (install the [claw] extra for multi-channel + subagent + cron features). <1k-token system prompt, 50+ LLM providers, AGENTS.md + skills context, session tree, prompt/auto permissions."
 readme = "README.md"
 requires-python = ">=3.12"

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/src/vtx/llm/provider.yaml

@@ -122,6 +122,7 @@ providers:
       - qwen/qwen3.7-max
       - z-ai/glm-5.2
       - nvidia/nemotron-3-ultra-550b-a55b:free
+      - tencent/hy3
     fetch_models: true
     models_endpoint: /models
     model_parser:
@@ -738,3 +739,15 @@ providers:
     models_endpoint: /models
     model_parser:
       cooldown_minutes: 60
+
+  - slug: meta
+    display_name: Meta AI
+    description: Meta AI API (Llama models).
+    family: openai_compat
+    base_url: https://api.meta.ai/v1
+    api_key_env: META_API_KEY
+    known_models: []
+    fetch_models: true
+    models_endpoint: /models
+    model_parser:
+      cooldown_minutes: 60

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/src/vtx/tools/__init__.py

@@ -1,3 +1,5 @@
+from typing import Any
+
 from ..core.types import ToolDefinition
 from .ask_user import AskUserTool
 from .background import BackgroundTaskManager, BackgroundTaskRecord, get_manager, set_manager
@@ -114,7 +116,7 @@ def get_tools_with_extensions(
 _SCHEMA_DROP_KEYS = frozenset({"title", "minLength", "maxLength", "minItems", "maxItems"})
 
 
-def _slim_schema(node: object) -> object:
+def _slim_schema(node: Any) -> Any:
     """Shrink pydantic JSON schema for the LLM: drop ``title`` and length
     constraints (still enforced by pydantic on the actual call) and collapse
     ``anyOf: [T, null]`` optional unions down to ``T``."""

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/src/vtx/version.py

@@ -39,7 +39,7 @@ else:
     try:
         VERSION = version(PACKAGE_NAME)
     except PackageNotFoundError:
-        VERSION = "0.2.4"  # Fallback version if package metadata is not available
+        VERSION = "0.2.5"  # Fallback version if package metadata is not available
 
 
 def format_version() -> str:

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/tests/test_config_migration.py

@@ -174,7 +174,7 @@ llm:
     reset_config()
     cfg = get_config()
 
-    assert cfg.llm.system_prompt.content.startswith("You are Vtx, an expert coding assistant.")
+    assert cfg.llm.system_prompt.content.startswith("You are Vtx, an expert coding agent.")
     assert "# Tool usage" not in cfg.llm.system_prompt.content
     assert "Old tool instruction" not in cfg.llm.system_prompt.content
     assert "~/.agents/skills" in cfg.llm.system_prompt.content
@@ -215,7 +215,7 @@ llm:
     cfg = get_config()
 
     assert cfg.llm.default_model == "legacy-model"
-    assert cfg.llm.system_prompt.content.startswith("You are Vtx, an expert coding assistant.")
+    assert cfg.llm.system_prompt.content.startswith("You are Vtx, an expert coding agent.")
     assert "legacy prompt" not in cfg.llm.system_prompt.content
     assert cfg.llm.system_prompt.git_context is True
 

{vtx_coding_agent-0.2.4 → vtx_coding_agent-0.2.5}/uv.lock

@@ -3376,7 +3376,7 @@ wheels = [
 
 [[package]]
 name = "vtx-coding-agent"
-version = "0.2.4"
+version = "0.2.5"
 source = { editable = "." }
 dependencies = [
     { name = "aiofiles" },