v0.0.24

• Restrict the depth and length of evaluation results

v0.0.23

• Avoid using potentially incorrectly auto-balanced code in evaluations and edits
• Fix: Insert and replace top level forms doesn't work on one-line forms ending in a line comment

v0.0.22

• Fix: Evaluation tool not available in v0.0.21

v0.0.21

• Make the structural editing tools return file context when the tool call is invalid
  • Add user settings/configurability: calva-backseat-driver.editor.fuzzyLineTargetingPadding and calva-backseat-driver.editor.lineContextResponsePadding

v0.0.20

• Update tool ids to qualify them as Clojure
• Remove Human Intelligence tool

v0.0.19

• Fix: The structural insert/replace top-level forms tools do not prevent editing non-top-level forms
• Add tool for creating files
• Add tool for appending top-level forms to a file

v0.0.18

• Remove default copilot instructions settings
• Improve instructions to agents for the structural editing tools

v0.0.17

• Balance brackets before evaluating code

v0.0.16

• Human Intelligence input box closes when focus is lost (should persist)

v0.0.15

• Limit the lines of REPL output to be passed to the AI by the REPL output tool