Skip to content

Releases: cecli-dev/cecli

v0.98.0

31 Mar 03:28
57029fa

Choose a tag to compare

What's Changed

MR: #472
Full Changelog: v0.97.5...v0.98.0

Addresses:
#469

Includes:

  • Add support for multiple --mcp-servers-files arguments by @szmania in #473
  • Add agent model display and configuration improvements by @szmania in #470
  • Modify ShowNumberedContext into ShowContext for content based targeting for stable hashline/hashpos based edits
  • Add "rules"-like file support to allow dynamically loadable files similar to what AGENTS.md or CLAUDE.md are used for
  • Add --rules argument and /rules command to support the above
  • Agent mode lints files as it edits them to detect syntax errors and ameliorate them during operation
  • Modify similarity detection so it is less intrusive
  • Change hashline/hashpos seperator since most LLM tokenizers see double colons as a single token (nice on the eyes, nice for the machine)
  • Conversation system decomposition into an instance-able set of base classes that will support multiple context streams for sub-agent development
  • Memory usage improvements by using GitPython's GitCmdObjectDB and general caching to search through the underlying git repository for generation of the repository map
  • Agent mode prompt updates to make the steps more easily accomplishable to the models and use slightly fewer tokens
  • Make available commands easier to search through by matching sub strings after start strings
  • Fix quoted file names being incorrectly spliced in the /add-like commands
  • Allow sending/injecting messages while a generation task is running in the TUI to increase general real-time steerability

0.97.5

17 Mar 17:51
f6a3bba

Choose a tag to compare

What's Changed

MR: #466
Full Changelog: v0.97.4...v0.97.5

Includes:

  • Updates to caching behavior so anthropic models cache prompts appropriately
  • Force output to use utf-8 encoding generally
  • Update to content addressing in hashline file representation
  • Agent mode updates to partial file content message ordering
  • Resize repomap at run time with /map-refresh <int>

v0.97.4

10 Mar 13:37
3f41a1b

Choose a tag to compare

What's Changed

MR: #464
Full Changelog: v0.97.3...v0.97.4

Addresses:

Includes:

  • Use the configured shell when tool command is called by @jasonbcox in #465
  • Auto create .cecli/workspace in agent mode so the agent has a place to put temp files
  • Updated context block management to support file inclusion/exclusion operations
  • Update model metadata
  • Prevent unneeded test commands from running
  • Allow Grep tool to work on single files
  • Allow wrapped lines in the TUI to be moved through vertically instead of automatically navigating between input prompts
  • Add --cost-limit parameter
  • Allow automatic parsing of XML-like tool calls attempted by some local models

New Contributors

v0.97.3

05 Mar 14:05
ebd79a4

Choose a tag to compare

What's Changed

MR: #461
Full Changelog: v0.97.2...v0.97.3

Addresses:
#460

Includes:

  • Add completions to /save-session command
  • Hot reloading skills when hot_reload is true under agent-config (useful to making/using skills at runtime)
  • New logic for response loop prevention in agent mode (output similarity and adaptive token penalties)
  • Fix --terminal-setup flag crashing on initialization

v0.97.2

03 Mar 07:32
49fc445

Choose a tag to compare

What's Changed

MR: #452
Full Changelog: v0.97.1...v0.97.2

Addresses:

Includes:

  • The entire hooks subsystem, add python file or command hooks at the following points:
    start, end, on_message, end_message, pre_tool, post_tool
  • Hashline edit format updates/optimizations
  • Clear file diff message on /undo command call
  • ShowNumberedContext tool context efficiency by merging requests into singular representation
  • /add respects --add-gitignore-files setting
  • #455 Make --watch-files work in dumb terminals and non-fancy input by @ekr1

New Contributors

  • @ekr1 made their first contribution in #455

v0.97.1

22 Feb 05:32
54bec5f

Choose a tag to compare

What's Changed

MR: #448
Full Changelog: v0.97.0...v0.97.1

Addresses:

Includes:

  • Add ability to set agent model in config and add in-chat command by @gopar in #444
  • New hashline edit format both standalone and as the vehicle for agent mode since it simplifies text replacement
  • Allow models with colons in name to have model overrides
  • Add model_settings key under overrides to modify model settings in the overrides section without needing to declare a model settings file

v0.97.0

14 Feb 17:44
94b82eb

Choose a tag to compare

What's Changed

MR: #438
Full Changelog: v0.96.10...v0.97.0

Addresses:

Includes:

  • task() method exported from cecli.main to run system programmatically as a part of another program
  • /terminal-setup updates to allow the user to confirm configuration edits
  • Fix JSON parsing from raw responses so weaker models can perform tool calling outside of the OpenAI tool calling schema
  • Automatic command backgrounding so agent cannot get stuck on miscalled, non-exiting command
  • New key bindings for TUI (input_start, input_end, output_up, and output_down)

v0.96.10

08 Feb 15:28
79c448c

Choose a tag to compare

What's Changed

MR: #435
Full Changelog: v0.96.9...v0.96.10

Addresses:

Includes:

  • Simplify line based editing tools (InsertText and DeleteText) to work off of pure line numbers
  • Make Grep tool capable of multi-searching
  • Speed up git ignored file checking by a factor of ~100
  • Speed up file mention checking due to the above
  • Small tool error mitigations

v0.96.9

03 Feb 13:55
3c64c9b

Choose a tag to compare

What's Changed

MR: #431
Full Changelog: v0.96.8...v0.96.9

Includes:

  • Fix for error on switching to gemini models from non-gemini model
  • Fix for error when model duplicates tool call responses

v0.96.8

02 Feb 05:35
d43775e

Choose a tag to compare

What's Changed

MR: #427
Full Changelog: v0.96.7...v0.96.8

Includes:

  • Format system messages and in context files before adding user message to conversation stream
  • Use relative paths consistently in conversation messages so models know we are referring to the same files