Accord v0.1.0-beta.1

Accord turns agent tasks into executable contracts and posts a proof-of-done verdict:

• what passed
• what failed
• what still needs human review

Install

pip install "accord @ git+https://github.com/JohnODowdAI/accord.git@v0.1.0-beta.1"

Best first use case

• one narrow task type (path-limited refactors, required tests, API compatibility)
• one small contract
• one real PR

Feedback I want

1. time to write the first contract
2. whether the summary was understandable immediately
3. whether it caught anything useful
4. whether needs_human felt helpful or annoying
5. whether you'd keep it for that task type

Intentionally deferred

• LLM semantic grading
• GitHub App / hosted checks
• plugin system
• dashboard