Full Changelog: v0.3.1...v0.3.2

Full Changelog: v0.3.0...v0.3.1

v0.2.0

• New Rust-based TUI architecture
• Fully interactive AI agent execution
• Tool-driven command system
• Streaming logs and execution control

Full Changelog: v0.1.4...v0.2.0

OsmoGrep v0.1.4 – Stability, Sandboxing & Parallel Execution

🚀 What’s New

1. Improved diff analysis stability

• Hardened extract_full_function logic
• More accurate multi-line function boundary detection
• Fewer edge-case failures and zero panics in irregular diffs

2. Fixed TextDiff import issues

• Removed deprecated utils::diff usage
• Clean InlineChange handling
• Confirmed compatibility with latest similar crate

3. Sandboxed execution

• Fully isolated temp directories for deterministic test runs
• Eliminates FileNotFound errors from shared state
• Improved error reporting in deeply nested commands

4. Parallel execution support

• Multiple test commands can now execute concurrently
• Faster validation cycles when scanning larger codebases
• Built-in guardrails to avoid race conditions and temp-file clashes

5. General cleanup

• Reduced noisy stdout logs
• Minor performance improvements in diff scanning
• Internal refactors + version bump

Summary

v0.1.4 delivers stronger stability, safer execution, and new parallel test handling, making OsmoGrep more reliable for autonomous validation workflows and future Greptile agent integration.

Full Changelog: v0.1.4...v0.1.4

OsmoGrep v0.1.1

This is the first public release of OsmoGrep, a terminal-native execution agent designed for safe, diff-scoped test generation and execution within local repositories.

What’s included

Core agent model

• Diff-scoped analysis of uncommitted changes
• Explicit agent branches for isolated execution
• No implicit git mutations or auto-applied changes
• Human-in-the-loop execution model

Context system

• Full context snapshot built from diffs and repository tests
• AST-based symbol and import extraction (Python and Rust)
• Context graph surfaced explicitly to the agent

Test generation & execution

• LLM-driven single-test generation with retry and feedback loop
• Deterministic semantic cache for generated tests
• Safe materialization and execution of generated tests
• Full test suite execution with structured reporting

Reporting

• Markdown test suite reports
• Raw logs and structured index artifacts
• Failure extraction with file and test attribution

Terminal UI

• Interactive TUI with command input, logs, and diff view
• Explicit execution state and lifecycle visibility
• Inspector for diffs, risks, and detected symbols

LLM backends

• Local inference via Ollama
• Remote inference via OpenAI-compatible APIs
• Prompt caching and reload controls

Installation

• Single-line installer via hosted binary
• No cloning or manual build required

Intended use

OsmoGrep is designed as a deterministic, test-centric execution agent for validating code changes safely on active working trees.
It is not a general-purpose autonomous coding agent.

Notes

• Linux x86_64 binary included
• macOS and additional architectures planned
• API and internals may change rapidly in early releases

Full Changelog: https://github.com/kaushal07wick/OsmoGrep/compare/v0.1.1

OsmoGrep v0.1.0

This is the first public release of OsmoGrep, a terminal-native execution agent designed for safe, diff-scoped test generation and execution within local repositories.

What’s included

Core agent model

• Diff-scoped analysis of uncommitted changes
• Explicit agent branches for isolated execution
• No implicit git mutations or auto-applied changes
• Human-in-the-loop execution model

Context system

• Full context snapshot built from diffs and repository tests
• AST-based symbol and import extraction (Python and Rust)
• Context graph surfaced explicitly to the agent

Test generation & execution

• LLM-driven single-test generation with retry and feedback loop
• Deterministic semantic cache for generated tests
• Safe materialization and execution of generated tests
• Full test suite execution with structured reporting

Reporting

• Markdown test suite reports
• Raw logs and structured index artifacts
• Failure extraction with file and test attribution

Terminal UI

• Interactive TUI with command input, logs, and diff view
• Explicit execution state and lifecycle visibility
• Inspector for diffs, risks, and detected symbols

LLM backends

• Local inference via Ollama
• Remote inference via OpenAI-compatible APIs
• Prompt caching and reload controls

Installation

• Single-line installer via hosted binary
• No cloning or manual build required

Intended use

OsmoGrep is designed as a deterministic, test-centric execution agent for validating code changes safely on active working trees.
It is not a general-purpose autonomous coding agent.

Notes

• Linux x86_64 binary included
• macOS and additional architectures planned
• API and internals may change rapidly in early releases

Full Changelog: https://github.com/kaushal07wick/OsmoGrep/commits/v0.1.0