v0.4.0 - Open Source Apache License

What's Changed

• Publish MCP Pipeline by @Dimwiddle in #108
• Add MCP publish workflow by @Dimwiddle in #109
• MCP Fix Publish pipeline by @Dimwiddle in #110
• SpecLeft becomes fully open source by @Dimwiddle in #120
• Bump python-frontmatter from 1.0.1 to 1.1.0 by @dependabot[bot] in #71
• Bump pyyaml from 6.0.2 to 6.0.3 by @dependabot[bot] in #118

Full Changelog: v0.3.0...v0.4.0

v0.3.0

What's Changed

• Local agent github workflow by @Dimwiddle in #96
• Align CLI version with package metadata (#87) by @Dimwiddle in #97
• Add SpecLeft skill file security and integrity workflow (#95) by @Dimwiddle in #98
• Optimize CLI output defaults for agent efficiency (#94) by @Dimwiddle in #99
• Add MCP server for agent discovery (#76) by @Dimwiddle in #100
• Implement v0.3.0 MCP security hardening (#91) by @Dimwiddle in #101
• Readme updates and MCP coverage fix by @Dimwiddle in #102
• 77 feature test specleft mcp and setup by @Dimwiddle in #107

Full Changelog: v0.2.2...v0.3.0

v0.2.2

What's Changed

• Fix skeleton/validate/stats not finding plan-generated specs by @Dimwiddle in #86

Full Changelog: v0.2.1...v0.2.2

v0.2.1

What's Changed

• Add PRD template and update plan command to use default template by @Dimwiddle in #84

Full Changelog: v0.2.0...v0.2.1

v0.2.0

🪜 Add Features and Scenarios Command — Incrementally add features and scenarios via the CLI for quick enhancement of a plan
🔖 Enhanced PRD parsing - More flexible converting of prd.md in to feature units.
📜 Logged Feature Changes - Keep a historical trace of features and scenarios added to the project. Provides externalised memory to agents
📖 Agent Guide - Provide clarity and guidance for agents to know how to best proceed in scenarios such as: refactoring, cleanup, regression bugs, features and scenarios.
🔄 Async test handling - Async tests are now supported with https://github.com/SpecLeft decorator and step context manager
🧪 Test Stubs - Create empty test containers as an alternate to test skeletons.

Deprecated

⚠️ The features/ directory will be migrated to .specleft/specs - existing projects using features will keep working, but I'd recommend you migrate the feature units in to .specleft/specs over the next few releases

v0.1.1

Patch

Modified src/specleft/commands/doctor.py: Updated the regex to r"[<>=\s[~!;@]". This ensures robust parsing of dependency names, handling extras, version constraints, environment markers, and URL references correctly.

The doctor command is now robust against complex dependency specifications in pyproject.toml.

v0.1.0

SpecLeft v0.1.0 — Initial Beta Release

This is the first public beta of SpecLeft, a planning-first CLI for humans and AI coding agents to externalize feature intent before writing code, with optional CI enforcement for Python projects.

What’s included

🧭 Planning-first workflow

• Convert a prd.md into feature specifications with:

  specleft plan

• Generates human-reviewable feature files under features/*.md
• No code or tests are written by default

🧩 Canonical feature format

• One feature per file
• Scenarios with explicit priorities
• Optional metadata, safe defaults, and predictable parsing

🤖 Agent-aware tooling

• Agent contract introspection:

  specleft contract --format json

• Designed to support autonomous, plan-first agent workflows

🧪 Test skeletons & progress tracking

• Generate test skeletons from declared behaviour
• Identify the next scenario to implement
• Track status and coverage without enforcing by default

🚦 Optional CI enforcement (beta)

• Enforce declared critical/high behaviour using a signed policy:

  specleft enforce <policy.yml>

• Calm, explanatory output intended for real CI pipelines

What this release is (and isn’t)

This is a beta.
The goal of v0.1.0 is to validate:

• agent adoption
• CI tolerance
• intent-first workflows in real projects

SpecLeft:

• does not replace pytest
• does not reinterpret your tests
• does not mutate code unless explicitly requested

License

SpecLeft is source-available under the PolyForm Shield License 1.0.0.
See LICENSE for details.

Feedback

Early feedback is very welcome — especially from:

• platform / principal engineers
• teams experimenting with AI-assisted development
• anyone running SpecLeft in CI

Issues and discussions are open in the repository.

OR

Email at richard@specleft.dev

---

Thanks for trying SpecLeft.