Skip to content

V0.98.0#472

Merged
dwash96 merged 56 commits intomainfrom
v0.98.0
Mar 31, 2026
Merged

V0.98.0#472
dwash96 merged 56 commits intomainfrom
v0.98.0

Conversation

@dwash96
Copy link
Copy Markdown
Collaborator

@dwash96 dwash96 commented Mar 27, 2026 •
edited
Loading

Addresses:
#469

Includes:

  • feat: Add support for multiple --mcp-servers-files arguments #473
  • feat: Add agent model display and configuration improvements #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

Your Name and others added 30 commits March 21, 2026 05:58
Bump Version
8055471
Update conversation system so most recent user or assistance message …
d52f9f7 
…is the last thing the LLM sees
Agent Mode Updates:
4185c20 
- ShowContext tool re-imagined to not work on line numbers in favor of patterns in the file in question
- Enforce inability to edit until ShowContext is called
- Add more obscurity to content bits of hashpos representation so LLM does not accidentally learn patterns in the ending bits
- Don't allow hashlines to approximate anywhere in the file, use 30 line cap (that maybe can become configurable) for how tolerant of content based non exact matches we are
Add rules file support (for loading things like AGENTS.md files and o…
9b45684 
…ther long persistent context, like project plans)

- /rules command
- --rules argument
- General support for token counting, context management, hot reloading
- Update when linting and edit restrictions occur to interrupt the LL…
e2aeebf 
…M less while fixing mistakes

- Remove indent_text tool since it almost never gets used
Add parameter to not show linter output in terminal
4de1255
Disable edit_allowed unless in the case of total edit failure
f41db8c
Give diff messages even higher priority, make similarity more stringe…
a4b2b34 
…nt to not interrupt the model as much
Experimentation with change targetting
feecd7a
Propagate errors even more again
bf6db4c
Update show_context.py
52cc33e
ShowContext messaging update
a21bd42
ShowContext: allow end pattern to be repetitive to get the closest match
064af76
Update deps
20b3bf0
@szmania
fix: AgentCoder uses main_model as fallback when agent_model not set
194b99c 
Co-authored-by: cecli (openai/openai_gemini_cli/gemini-2.5-pro)
@szmania
fix: Make AgentCoder.send an async generator
2cc1b15 
Co-authored-by: cecli (openai/openai_gemini_cli/gemini-2.5-pro)
@szmania
chore: Update agent model spinner text in TUI
f1f5d1e 
Co-authored-by: cecli (openai/openai_gemini_cli/gemini-2.5-pro)
@szmania
fix: Update spinner text to show active agent model
d694568 
Co-authored-by: cecli (openai/openai_gemini_cli/gemini-2.5-pro)
@szmania
fix: Update model name in footer when in agent mode
ddd7850 
Co-authored-by: cecli (openai/openai_gemini_cli/gemini-2.5-pro)
@szmania
fix: Update footer to display active model name
a7de988 
Co-authored-by: cecli (openai/openai_gemini_cli/gemini-2.5-pro)
@szmania
cli-2: fix formatting
72f3b68
@szmania
feat: Support multiple --mcp-servers-file arguments
2f8edbd 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
refactor: Update mcp/utils.py to handle multiple server files
a51c0d3 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
feat: Add test for multiple mcp-servers-file arguments
5beb691 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
feat: Add support for multiple --mcp-servers-file arguments
59c5166 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
feat: Add support for multiple --mcp-servers-file arguments
7307d76 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
feat: Add test for multiple mcp-servers-file arguments
f37c97f 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
feat: Add test for multiple mcp-servers-file arguments
7e1ade0 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
feat: Add support for multiple --mcp-servers-file arguments
1794eca 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
cli-3: multiple mcp servers files
7385512
szmania and others added 26 commits March 26, 2026 11:27
@szmania
feat: Add support for multiple --mcp-servers-files and deprecate old arg
5709853 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
@szmania
fix: Update tests/basic/test_main.py for argument and mock changes
d5f618a 
Co-authored-by: cecli (openai/gemini_cli/gemini-2.5-pro)
Make conversation classes instantiable, add common interface to acces…
e35c9cc 
…s conversation system per coder class as a nother step towards sub-agent functionality
Update hashpos format to delimit on double colons, saves tokens, does…
a02a6c8 
…n't generally harm accuracy
Modify repetition prompts to be less hostile to the LLM
65d6d69
@dwash96
Merge pull request #473 from szmania/cli-3-multiple-mcp-servers-files
6987cb4 
feat: Add support for multiple --mcp-servers-files arguments
Add memory profiling tools to dev package
f257286
Use GitCmdObjectDB instead of GitDB per GitPython's documentation
f47874b
Bump Version
3b714e5
@dwash96
Merge pull request #470 from szmania/cli-2-agent-model
e155562 
feat: Add agent model display and configuration improvements
Use get_active_model() instead of get_active_model_name() to fully ab…
a69f24d 
…stract model object access
Simplify model switching and state propagation with from_model argument
508c862
Extraneous updates
49c3cde
Preserve ending new lines
08bcd68
Update hooks tests to actually run
42efde4
Fix test testing the wrong thing
6b84025
#469: Add asyncio io cancellation exception for missing mcp servers
67e65dd
Preserve latest assistant turn during context compaction
122fb87
Update README.md
d1444cb
Allow sub word command completion
954514b
Update favicon in embedded links
c30de80
Make sure quoting file names with spaces in them works for the /add-l…
465da6b 
…ike commands
Make sure command tools respect context-management-enabled for respon…
b7f1db2 
…se truncation
#442: Update published leaderboard, general documentation updates
fb7481c
Update TUI presentation
26d56b3
Allow the injection of messages into the coder class while the tui is…
f23240d 
… running a generation task for greater steerability
@dwash96 dwash96 merged commit 57029fa into main Mar 31, 2026
13 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@dwash96 @szmania