Changelog

1.0.1 (2026-11-08)

Fixes to the approval workflow, documentation rendering, and Python support policy. No change to the blueprint output format.

Fixed

• A rejected approval gate now stays awaiting approval. The skill revises that layer and presents it again, and does not start the next step until it is approved. If the user cannot be reached, it proceeds on defaults and flags the gate for review.
• The report step and final checklist state explicitly that nothing is committed unless the user asks.
• The verify step runs evals/grade.py only when it exists, using the command documented in the procedure.
• A broken table row in the SKILL.md file index is fixed.
• The --budget help text in evals/grade.py now points to references/vocabulary.md and gives the correct budget guidance.

Changed

• Minimum supported Python is now 3.12. The CI matrix tests 3.12 and 3.13.
• The skill description is shorter and easier to scan.
• The per-agent output line names the four supported targets: Copilot, Claude Code, Cursor, and AGENTS.md.
• The § symbol is replaced with the word "section" throughout the references, templates, and Shortlink example. The traceability register generator is unaffected.

Documentation

• The README has an Origin section explaining how the skill came about, and a Pantry section crediting EARS, Architecture Decision Records, and Anthropic's skill-creator.

1.0.0 (2026-11-07)

Initial release of mise-en-place, a spec-driven planning skill that turns requirements into a traceable project blueprint before implementation begins.

Highlights

• Supports new projects, existing codebases, scope changes, and resuming an incomplete blueprint.
• Produces structured intent, EARS requirements, architecture and ADRs, a phased roadmap and backlog, validation guidance, and a traceability register.
• Generates coding-agent customisations to keep implementation aligned with the approved blueprint.
• Available as a plugin for GitHub Copilot CLI, Claude Code, and Cursor, or as a skill copied into another supported agent environment.
• Includes standard-library scripts for register generation and link checking, a complete Shortlink example, and evaluation scenarios.