aia 2.0.0 → 2.1.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: 0f3c92628b523cd32fa77fa52f1546f1ec43ee9b65c645b33adb508086ef0c13
4
- data.tar.gz: cc9e9aab97552cff87e4b05b001d8330a99a5fea9000c6d2b76f07b1e1a7054e
3
+ metadata.gz: 1d7d1dff13dc9c4e4558c4624bf6368cb32f42a4a4896238e1750e5beaa24d7b
4
+ data.tar.gz: e430f69d905314342a538d063b5123dd5fd454dabcd6bae08a2ad3d696d74d56
5
5
  SHA512:
6
- metadata.gz: abf1b9d9f077b6d8aadd6c259e1cf3125007718b7912be857003f4189d57be4c20451093e87c6ec47f0a3be06c4ddf7cf767ffff5dc55be4177241e2d6fa8d7d
7
- data.tar.gz: 14e85d72e471d1bf8572d8fb2b417a6e1e0c641bc34e376b8748dd8b2f55665221fa6ff78966692f0ecce65ee990067c65a898db0d96f718ad713ec8f02048ec
6
+ metadata.gz: b136ae2840a452e8aa5dae4585588036ea10420abbd071c5fd3b45e0fea86a2f6c35005680757e8c2ebde9a63d8a6abf3538a164446c03cc1300aafd41a97728
7
+ data.tar.gz: 06c4e7081bdfe6c618bb4dc868d37daf232c491d04319bf8426232b85442a94eb279b04cba69e3e456d7580d51bdaede5d761367f9331f481387c15432506df4
data/CHANGELOG.md CHANGED
@@ -1,6 +1,49 @@
1
1
  # Changelog
2
2
 
3
- ## [2.0.0] 2026-11-02
3
+ ## [2.1.0] - 2026-11-07
4
+
5
+ Coding-agent CLIs (Claude Code, Codex, GitHub Copilot) can be used as models, and a prompt file can ask for a special mode such as `debate` in its front matter. Batch runs now exit 1 when an LLM call fails. The unit test suite runs without API keys, a UTF-8 locale or a particular Ruby patch level, and every test file runs on its own.
6
+
7
+ ### Added
8
+
9
+ - **Coding agents as models.** `--model claude_code/opus`, `codex/default` or `copilot/auto` sends prompts to the Claude Code, Codex or GitHub Copilot CLI in batch mode (stdin in, JSON out) through the `ruby_llm-providers-claude_code`, `-codex` and `-copilot` gems, which resume the agent's own session between turns. A bare provider name (`--model claude_code`) now means that provider's default model. Providers that answer `accepts_tools?` with `false` get no aia tools or MCP servers attached, with a one-line notice. An installed provider gem is found and loaded even when aia itself runs under Bundler (a `bin/aia` checkout, or an app that lists aia in its Gemfile) and the gem is not in that Gemfile, provided the bundle already satisfies the gem's dependencies.
10
+
11
+ - **`mode:` front matter key.** A prompt file can ask for one of the special modes that the chat directives `/debate`, `/verify`, `/decompose`, `/delegate`, `/spawn`, `/orchestrate` and `/concurrent` request, by naming it in its front matter: `mode: debate`. Prompts run from the command line, in a pipeline or as executable prompts then get that mode, which until now ran only in chat; a `/debate` line inside a prompt file was sent to the model as text. As in chat, a debate needs several robots (`--model a,b,c`).
12
+
13
+ ### Fixed
14
+
15
+ - **`--provider` outranks `AIA_MODEL` and the config file.** `--provider claude_code` replaced the provider only of models given without a `<provider>/` prefix, so with `AIA_MODEL=lms/qwen3.8` exported the flag did nothing and the prompt still went to LM Studio. A command-line `--provider` now applies to every model that came from the environment or a config file, prefix or not, keeping their model names; it sets the provider, never the model, so `--model claude_code` (a bare provider name) remains the way to ask for a provider's default model. Models named on the command line with `--model` keep an explicit prefix, since both values come from the same place. When a provider gem cannot be loaded because aia is running under Bundler, the warning now names the Gemfile to add the `ruby_llm-providers-<slug>` gem to.
16
+
17
+ - **Batch runs exit 1 on LLM errors.** robot_lab shields a robot's exception into an `Error: ...` reply so one failing robot cannot crash a network, which left `aia prompt` printing a bad API key, rate limit or provider outage as the answer and exiting 0. A hook now records the real exception for each pipeline step: outside chat the step fails, the error goes to stderr and aia exits 1; in a chat session the error is shown and the session continues. Unshielded RubyLLM errors (image generation, provider setup) exit 1 with a one-line message instead of a backtrace.
18
+
19
+ - **`--paint` works again, and `--image_quality` / `--image_style` reach the API.** Since 2.0.0 the image step handed `RubyLLM.paint` a `params:` keyword that ruby_llm 2.1 does not have, so every `--paint` run stopped with `ArgumentError: unknown keyword: :params` before any request was made. The options now go through `provider_options:`, which ruby_llm merges into the request, so `--image_quality hd --image_style vivid` take effect (#135).
20
+
21
+ - **`DelegateHandler` and `DelegateToForemanTool` load what they use.** `delegate_handler.rb` now requires `TaskDecomposer` and `TaskExecutor`, and `delegate_to_foreman_tool.rb` requires `TaskBoardTool`. Each worked only because its one caller also required those files, so loading either on its own and calling it raised `uninitialized constant`.
22
+
23
+ **Test suite**
24
+
25
+ - **Failing test runs exit non-zero.** `test_helper.rb`'s `exit` override, which stops aia's exit calls from ending a test run early, also swallowed Minitest's exit status, so `rake test` and single test files exited 0 when tests failed. The override now stands down after the suite runs.
26
+
27
+ - **Test suite loads on Ruby 4.1.** `ostruct` is a bundled gem since Ruby 4.0 and only reached the test bundle as a transitive dependency of `oj` 3.17.7; `oj` 3.17.8 dropped it, so `bundle update` left the suite unable to load `OpenStruct`. It is now a development dependency. The speech player test no longer stubs `ENV`, which Ruby 4.1 rejects as a shareable object.
28
+
29
+ - **Unit suite passes without API keys or a UTF-8 locale.** The network-building tests create real RubyLLM chats, which ruby_llm 2.1 refuses without a provider API key, so 37 tests errored unless `OPENAI_API_KEY` was set. They now skip that credential check; they never call an API. The gemspec version test errored under a non-UTF-8 locale; it now loads the gemspec and compares its version with `AIA::VERSION`. `test_helper.rb` requires `ostruct`, so test files that use `OpenStruct` also run on their own.
30
+
31
+ - **Every unit test file runs on its own.** Six files raised `uninitialized constant` when run alone, because aia loads a class each one uses lazily: `TaskDecomposer`, `TaskExecutor`, `TaskBoardTool` (one file each) and `ToolFilter::TFIDF` (three files). Each file now requires that class.
32
+
33
+ - **`ConfigTest` ignores the shell's `AIA_PLUGINS_DIR`.** The test that checks `paths.plugins_dir` defaults to `nil` read whatever the developer's shell had exported and failed, and the two tests that set the variable deleted it afterwards. All three now set the variable for their own run and restore it.
34
+
35
+ ## [2.0.1] - 2026-11-02
36
+
37
+ ### Fixed
38
+
39
+ - **`--chat` startup output is readable again.** ruby_llm 2.1 starts stdio MCP servers with their stderr connected to aia's, so server banners, npm notices, stack traces and validation dumps were printed over the "Connecting MCP servers" spinner and garbled it. That stderr now goes to `~/.config/aia/logs/mcp_servers.log`. Set `logger.mcp.server_stderr` to another path, `STDERR` (the 2.0.0 behavior) or `null` to discard it.
40
+
41
+ ### Changed
42
+
43
+ - `logger.mcp.file` defaults to `~/.config/aia/logs/mcp.log` (was `STDOUT`), so MCP connection warnings no longer print over the spinner. Failed servers are still listed in the startup banner. A config file that sets `file: STDOUT` keeps the old behavior.
44
+ - Log directories are created when they don't exist.
45
+
46
+ ## [2.0.0] - 2026-11-02
4
47
 
5
48
  AIA 2.0 rebuilds the CLI on [robot_lab](https://github.com/MadBomber/robot_lab): every session is a crew of robots, MCP servers come from ruby_llm 2.1's built-in client, and any provider RubyLLM registers can be used. This entry summarizes everything since 1.1.1, the last 1.x release, including the `2.0.0.pre.alpha` and `2.0.0.pre.beta2` pre-releases.
6
49
 
@@ -425,7 +468,7 @@ If you used a 2.0 pre-release: the KBS rule engine, `/rules`, user-defined rule
425
468
  - Added `handle_tool_crash` method that logs errors with 5-line traceback
426
469
  - Tool crashes no longer crash AIA - conversation continues gracefully
427
470
 
428
- ## [0.9.24] 2025-12-17
471
+ ## [0.9.24] - 2025-12-17
429
472
  ### Fixes
430
473
  - Ran into a problem with the `shared_tools` gem and the --require parameter of AIA which required changes to both gems.
431
474
 
@@ -450,7 +493,7 @@ If you used a 2.0 pre-release: the KBS rule engine, `/rules`, user-defined rule
450
493
  - Added `handle_tool_crash` method for graceful error handling
451
494
  - Removed debug statements
452
495
 
453
- ## [0.9.23] 2025-12-06
496
+ ## [0.9.23] - 2025-12-06
454
497
 
455
498
  ### New Features
456
499
  - **MCP Server Configuration**: Added native support for defining MCP (Model Context Protocol) servers in the config file
@@ -483,7 +526,7 @@ If you used a 2.0 pre-release: the KBS rule engine, `/rules`, user-defined rule
483
526
  MY_VAR: "value"
484
527
  ```
485
528
 
486
- ## [0.9.22] 2025-11-12
529
+ ## [0.9.22] - 2025-11-12
487
530
 
488
531
  ### Bug Fixes
489
532
  - **TEST SUITE**: Fixed all Mocha test isolation issues causing stub contamination between tests
@@ -499,14 +542,14 @@ If you used a 2.0 pre-release: the KBS rule engine, `/rules`, user-defined rule
499
542
  - Standardized stub usage pattern across test suite for consistency
500
543
  - Eliminated stub leakage that caused "unexpected invocation" and "AIA was instantiated in one test but receiving invocations in another" errors
501
544
 
502
- ## [0.9.21] 2025-10-08
545
+ ## [0.9.21] - 2025-10-08
503
546
  ### Bug Fixes
504
547
  - **Checkpoint Directive Output**: Fixed `//checkpoint` directive to return empty string instead of status message (lib/aia/directives/configuration.rb:155)
505
548
  - Prevents checkpoint creation messages from entering AI context
506
549
  - Outputs confirmation to STDOUT instead for user feedback
507
550
  - Prevents potential AI manipulation of checkpoint system
508
551
 
509
- ## [0.9.20] 2025-10-06
552
+ ## [0.9.20] - 2025-10-06
510
553
  ### Added
511
554
  - **Enhanced Multi-Model Role System (ADR-005)**: Implemented per-model role assignment with inline syntax
512
555
  - New inline syntax: `--model MODEL[=ROLE][,MODEL[=ROLE]]...`
@@ -570,7 +613,7 @@ If you used a 2.0 pre-release: the KBS rule engine, `/rules`, user-defined rule
570
613
  - Enhanced `apply_file_config_to_struct` to detect and process model arrays with role hashes
571
614
  - Enhanced `envar_options` to parse inline syntax for `:model` key (lib/aia/config/base.rb:212-217)
572
615
 
573
- ## [0.9.19] 2025-10-06
616
+ ## [0.9.19] - 2025-10-06
574
617
 
575
618
  ### Bug Fixes
576
619
  - **CRITICAL BUG FIX**: Fixed multi-model cross-talk issue (#118) where models could see each other's conversation history
@@ -633,7 +676,7 @@ bin/aia --chat --model lms/openai/gpt-oss-20b,ollama/gpt-oss:20b --tokens
633
676
  # (Previously: LMS would say "5 times", Ollama "6 times" due to cross-talk)
634
677
  ```
635
678
 
636
- ## [0.9.18] 2025-10-05
679
+ ## [0.9.18] - 2025-10-05
637
680
 
638
681
  ### Bug Fixes
639
682
  - **BUG FIX**: Fixed RubyLLM provider error parsing to handle both OpenAI and LM Studio error formats
@@ -652,7 +695,7 @@ bin/aia --chat --model lms/openai/gpt-oss-20b,ollama/gpt-oss:20b --tokens
652
695
  - Cleaned up debug puts statements from RubyLLMAdapter and provider_fix
653
696
  - Added robust JSON parsing with fallback error handling
654
697
 
655
- ## [0.9.17] 2025-10-04
698
+ ## [0.9.17] - 2025-10-04
656
699
 
657
700
  ### New Features
658
701
  - **NEW FEATURE**: Enhanced local model support with comprehensive validation and error handling
@@ -702,7 +745,7 @@ aia --model ollama/llama3.2 --chat
702
745
  > //models
703
746
  ```
704
747
 
705
- ## [0.9.16] 2025-09-26
748
+ ## [0.9.16] - 2025-09-26
706
749
 
707
750
  ### New Features
708
751
  - **NEW FEATURE**: Added support for Ollama AI provider
@@ -713,7 +756,7 @@ aia --model ollama/llama3.2 --chat
713
756
  - **ENHANCEMENT**: Expanded AI provider ecosystem with three new local/self-hosted model options
714
757
  - **ENHANCEMENT**: Improved flexibility for users preferring local LLM deployments
715
758
 
716
- ## [0.9.15] 2025-09-21
759
+ ## [0.9.15] - 2025-09-21
717
760
 
718
761
  ### New Features
719
762
  - **NEW FEATURE**: Added `//paste` directive to insert clipboard contents into prompts
@@ -723,7 +766,7 @@ aia --model ollama/llama3.2 --chat
723
766
  - Enhanced DirectiveProcessor with clipboard integration using the clipboard gem
724
767
  - Added comprehensive test coverage for paste directive functionality
725
768
 
726
- ## [0.9.14] 2025-09-19
769
+ ## [0.9.14] - 2025-09-19
727
770
 
728
771
  ### New Features
729
772
  - **NEW FEATURE**: Added `//checkpoint` directive to create named snapshots of conversation context
@@ -749,7 +792,7 @@ aia --model ollama/llama3.2 --chat
749
792
  - Refactored help directive to collect directives from all registered modules
750
793
  - Added comprehensive test coverage for checkpoint and restore functionality
751
794
 
752
- ## [0.9.13] 2025-09-02
795
+ ## [0.9.13] - 2025-09-02
753
796
  ### New Features
754
797
  - **NEW FEATURE**: Added `--tokens` flag to show token counts for each model
755
798
  - **NEW FEATURE**: Added `--cost` flag to enable cost estimation for each model
@@ -767,7 +810,7 @@ aia --model ollama/llama3.2 --chat
767
810
  - Enhanced test suite with improved assertions and coverage
768
811
  - Updated various gem dependencies to latest stable versions
769
812
 
770
- ## [0.9.12] 2025-08-28
813
+ ## [0.9.12] - 2025-08-28
771
814
 
772
815
  ### New Features
773
816
  - **MAJOR NEW FEATURE**: Multi-model support - specify multiple AI models simultaneously with comma-separated syntax
@@ -834,15 +877,15 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
834
877
  - TODO: focus on log file consistency using Logger
835
878
 
836
879
 
837
- ## [0.9.11] 2025-07-31
880
+ ## [0.9.11] - 2025-07-31
838
881
  - added a cost per 1 million input tokens to available_models query output
839
882
  - updated ruby_llm to version 1.4.0
840
883
  - updated all other gem dependencies to their latest versions
841
884
 
842
- ## [0.9.10] 2025-07-18
885
+ ## [0.9.10] - 2025-07-18
843
886
  - updated ruby_llm-mcp to version 0.6.1 which solves problems with MCP tools not being installed
844
887
 
845
- ## [0.9.9] 2025-07-10
888
+ ## [0.9.9] - 2025-07-10
846
889
  - refactored the Session and Config classes into more testable method_missing
847
890
  - updated the test suire for both the Session and Config classes
848
891
  - added support for MCP servers coming into AIA via the shared_tools gem
@@ -853,13 +896,13 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
853
896
  - //available_models now has context window size and capabilities for each model returned
854
897
 
855
898
 
856
- ## [0.9.8] 2025-06-25
899
+ ## [0.9.8] - 2025-06-25
857
900
  - fixing an issue with pipelined prompts
858
901
  - now showing the complete modality of the model on the processing line.
859
902
  - changed -p option from prompts_dir to pipeline
860
903
  - found problem with simple cov and deep cov w/r/t their reported test coverage; they have problems with heredoc and complex conditionals.
861
904
 
862
- ## [0.9.7] 2025-06-20
905
+ ## [0.9.7] - 2025-06-20
863
906
 
864
907
  - **NEW FEATURE**: Added `--available_models` CLI option to list all available AI models
865
908
  - **NEW FEATURE**: Added `//tools` to show a list of available tools and their description
@@ -870,7 +913,7 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
870
913
  - **DOCUMENTATION**: Updated README for better clarity and structure
871
914
  - **DEPENDENCY**: Updated Gemfile.lock with latest dependency versions
872
915
 
873
- ## [0.9.6] 2025-06-13
916
+ ## [0.9.6] - 2025-06-13
874
917
  - fixed issue 84 with the //llms directive
875
918
  - changed the monkey patch to the RubyLLM::Model::Modalities class at the suggestions of the RubyLLM author in prep for a PR against that gem.
876
919
  - added the shared_tools gem - need to add examples on how to use it with the --tools option
@@ -878,11 +921,11 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
878
921
  - added images/aia.png to README.md
879
922
  - let claude code rewrite the README.md file. Some details were dropped but overall in reads better. Need to add the details to a wiki or other documentation site.
880
923
 
881
- ## [0.9.5] 2025-06-04
924
+ ## [0.9.5] - 2025-06-04
882
925
  - changed the RubyLLM::Modalities class to use method_missing for mode query
883
926
  - hunting for why the //llms query directive is not finding image_to_image LLMs.
884
927
 
885
- ## [0.9.4] 2025-06-03
928
+ ## [0.9.4] - 2025-06-03
886
929
  - using RubyLLM v1.3.0
887
930
  - setting up a docs infrastructure to behave like the ruby_llm gem's guides side
888
931
  - fixed bug in the text-to-image workflow
@@ -890,7 +933,7 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
890
933
  - need to pay attention to the test suite
891
934
  - also need to ensure the non text2text modes are working
892
935
 
893
- ## [0.9.3rc1] 2025-05-24
936
+ ## [0.9.3rc1] - 2025-05-24
894
937
  - using ruby_llm v1.3.0rc1
895
938
  - added a models database refresh based on integer days interval with the --refresh option
896
939
  - config file now has a "last_refresh" String in format YYYY-MM-DD
@@ -899,41 +942,41 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
899
942
  - fixed a bug in the prompt_manager gem which is now at v0.5.5
900
943
 
901
944
 
902
- ## [0.9.2] 2025-05-18
945
+ ## [0.9.2] - 2025-05-18
903
946
  - removing the MCP experiment
904
947
  - adding support for RubyLLM::Tool usage in place of the MCP stuff
905
948
  - updated prompt_manager to v0.5.4 which fixed shell integration problem
906
949
 
907
- ## [0.9.1] 2025-05-16
950
+ ## [0.9.1] - 2025-05-16
908
951
  - rethink MCP approach in favor of just RubyLLM::Tool
909
952
  - fixed problem with //clear
910
953
  - fixed a problem with a priming prompt in a chat loop
911
954
 
912
- ## [0.9.0] 2025-05-13
955
+ ## [0.9.0] - 2025-05-13
913
956
  - Adding experimental MCP Client support
914
957
  - removed the CLI options --erb and --shell but kept them in the config file with a default of true for both
915
958
 
916
- ## [0.8.6] 2025-04-23
959
+ ## [0.8.6] - 2025-04-23
917
960
  - Added a client adapter for the ruby_llm gem
918
961
  - Added the adapter config item and the --adapter option to select at runtime which client to use ai_client or ruby_llm
919
962
 
920
- ## [0.8.5] 2025-04-19
963
+ ## [0.8.5] - 2025-04-19
921
964
  - documentation updates
922
965
  - integrated the https://pure.md web service for inserting web pages into the context window
923
966
  - //include http?://example.com/stuff
924
967
  - //webpage http?://example.com/stuff
925
968
 
926
- ## [0.8.2] 2025-04-18
969
+ ## [0.8.2] - 2025-04-18
927
970
  - fixed problems with pre-loaded context and chat repl
928
971
  - piped content into `aia --chat` is now a part of the context/instructions
929
972
  - content via "aia --chat < some_file" is added to the context/instructions
930
973
  - `aia --chat context_file.txt context_file2.txt` now works
931
974
  - `aia --chat prompt_id context)file.txt` also works
932
975
 
933
- ## [0.8.1] 2025-04-17
976
+ ## [0.8.1] - 2025-04-17
934
977
  - bumped version to 0.8.1 after correcting merge conflicts
935
978
 
936
- ## [0.8.0] WIP - 2025-04-15
979
+ ## [0.8.0] - 2025-04-15
937
980
  - Updated PromptManager to v0.5.1 which has some of the functionality that was originally developed in the AIA.
938
981
  - Enhanced README.md to include a comprehensive table of configuration options with defaults and associated environment variables.
939
982
  - Added a note in README.md about the expandability of configuration options from a config file for dynamic prompt generation.
@@ -941,7 +984,7 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
941
984
  - Ensured version consistency across `.version`, `aia.gemspec`, and `lib/aia/version.rb`.
942
985
  - Verified and updated documentation to ensure readiness for release on RubyGems.org.
943
986
 
944
- ## [0.7.1] WIP - 2025-03-22
987
+ ## [0.7.1] - 2025-03-22
945
988
  - Added `UIPresenter` class for handling user interface presentation.
946
989
  - Added `DirectiveProcessor` class for processing chat-based directives.
947
990
  - Added `HistoryManager` class for managing conversation and variable history.
@@ -955,7 +998,7 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
955
998
  - Improved error handling and user feedback for directive processing.
956
999
  - Enhanced logging and output options for chat sessions.
957
1000
 
958
- ## [0.7.0] WIP - 2025-03-17
1001
+ ## [0.7.0] - 2025-03-17
959
1002
  - Major code refactoring for better organization and maintainability:
960
1003
  - Extracted `DirectiveProcessor` class to handle chat-based directives
961
1004
  - Extracted `HistoryManager` class for conversation and variable history management
@@ -966,13 +1009,13 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
966
1009
  - Improved output handling to suppress STDOUT when chat mode is off and output file is specified
967
1010
  - Updated spinner format in the process_prompt method for better user experience
968
1011
 
969
- ## [0.6.?] WIP
1012
+ ## [0.6.?] - WIP
970
1013
  - Implemented Tony Stark's Clean Slate Protocol on the develop branch
971
1014
 
972
- ## [0.5.17] 2024-05-17
1015
+ ## [0.5.17] - 2024-05-17
973
1016
  - removed replaced `semver` with `versionaire`
974
1017
 
975
- ## [0.5.16] 2024-04-02
1018
+ ## [0.5.16] - 2024-04-02
976
1019
  - fixed prompt pipelines
977
1020
  - added //next and //pipeline directives as shortcuts to //config [next,pipeline]
978
1021
  - Added new backend "client" as an internal OpenAI client
@@ -981,82 +1024,82 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
981
1024
  - Added --voice default: alloy (if "siri" and Mac? then uses cli tool "say")
982
1025
  - Added --image_size and --image_quality (--is --iq)
983
1026
 
984
- ## [0.5.15] 2024-03-30
1027
+ ## [0.5.15] - 2024-03-30
985
1028
  - Added the ability to accept piped in text to be appended to the end of the prompt text: curl $URL | aia ad_hoc
986
1029
  - Fixed bugs with entering directives as follow-up prompts during a chat session
987
1030
 
988
- ## [0.5.14] 2024-03-09
1031
+ ## [0.5.14] - 2024-03-09
989
1032
  - Directly access OpenAI to do text to speech when using the `--speak` option
990
1033
  - Added --voice to specify which voice to use
991
1034
  - Added --speech_model to specify which TTS model to use
992
1035
 
993
- ## [0.5.13] 2024-03-03
1036
+ ## [0.5.13] - 2024-03-03
994
1037
  - Added CLI-utility `llm` as a backend processor
995
1038
 
996
- ## [0.5.12] 2024-02-24
1039
+ ## [0.5.12] - 2024-02-24
997
1040
  - Happy Birthday Ruby!
998
1041
  - Added --next CLI option
999
1042
  - Added --pipeline CLI option
1000
1043
 
1001
- ## [0.5.11] 2024-02-18
1044
+ ## [0.5.11] - 2024-02-18
1002
1045
  - allow directives to return information that is inserted into the prompt text
1003
1046
  - added //shell command directive
1004
1047
  - added //ruby ruby_code directive
1005
1048
  - added //include path_to_file directive
1006
1049
 
1007
- ## [0.5.10] 2024-02-03
1050
+ ## [0.5.10] - 2024-02-03
1008
1051
  - Added --roles_dir to isolate roles from other prompts if desired
1009
1052
  - Changed --prompts to --prompts_dir to be consistent
1010
1053
  - Refactored common fzf usage into its own tool class
1011
1054
 
1012
- ## [0.5.9] 2024-02-01
1055
+ ## [0.5.9] - 2024-02-01
1013
1056
  - Added a "I'm working" spinner thing when "--verbose" is used as an indication that the backend is in the process of composing its response to the prompt.
1014
1057
 
1015
- ## [0.5.8] 2024-01-17
1058
+ ## [0.5.8] - 2024-01-17
1016
1059
  - Changed the behavior of the --dump option. It must now be followed by path/to/file.ext where ext is a supported config file format: yml, yaml, toml
1017
1060
 
1018
- ## [0.5.7] 2024-01-15
1061
+ ## [0.5.7] - 2024-01-15
1019
1062
  - Added ERB processing to the config_file
1020
1063
 
1021
- ## [0.5.6] 2024-01-15
1064
+ ## [0.5.6] - 2024-01-15
1022
1065
  - Adding processing for directives, shell integration and erb to the follow up prompt in a chat session
1023
1066
  - some code refactoring.
1024
1067
 
1025
- ## [0.5.3] 2024-01-14
1068
+ ## [0.5.3] - 2024-01-14
1026
1069
  - adding ability to render markdown to the terminal using the "glow" CLI utility
1027
1070
 
1028
- ## [0.5.2] 2024-01-13
1071
+ ## [0.5.2] - 2024-01-13
1029
1072
  - wrap response when its going to the terminal
1030
1073
 
1031
- ## [0.5.1] 2024-01-12
1074
+ ## [0.5.1] - 2024-01-12
1032
1075
  - removed a wicked puts "loaded" statement
1033
1076
  - fixed missed code when the options were changed to --out_file and --log_file
1034
1077
  - fixed completion functions by updating $PROMPT_DIR to $AIA_PROMPTS_DIR to match the documentation.
1035
1078
 
1036
- ## [0.5.0] 2024-01-05
1079
+ ## [0.5.0] - 2024-01-05
1037
1080
  - breaking changes:
1038
1081
  - changed `--config` to `--config_file`
1039
1082
  - changed `--env` to `--shell`
1040
1083
  - changed `--output` to `--out_file`
1041
1084
  - changed default `out_file` to `STDOUT`
1042
1085
 
1043
- ## [0.4.3] 2023-12-31
1086
+ ## [0.4.3] - 2023-12-31
1044
1087
  - added --env to process embedded system environment variables and shell commands within a prompt.
1045
1088
  - added --erb to process Embedded RuBy within a prompt because have embedded shell commands will only get you in a trouble. Having ERB will really get you into trouble. Remember the simple prompt is usually the best prompt.
1046
1089
 
1047
- ## [0.4.2] 2023-12-31
1090
+ ## [0.4.2] - 2023-12-31
1048
1091
  - added the --role CLI option to pre-pend a "role" prompt to the front of a primary prompt.
1049
1092
 
1050
- ## [0.4.1] 2023-12-31
1093
+ ## [0.4.1] - 2023-12-31
1051
1094
  - added a chat mode
1052
1095
  - prompt directives now supported
1053
1096
  - version bumped to match the `prompt_manager` gem
1054
1097
 
1055
- ## [0.3.20] 2023-12-28
1098
+ ## [0.3.20] - 2023-12-28
1056
1099
  - added work around to issue with multiple context files going to the `mods` backend
1057
1100
  - added shellwords gem to sanitize prompt text on the command line
1058
1101
 
1059
- ## [0.3.19] 2023-12-26
1102
+ ## [0.3.19] - 2023-12-26
1060
1103
  - major code refactoring.
1061
1104
  - supports config files \*.yml, \*.yaml and \*.toml
1062
1105
  - usage implemented as a man page. --help will display the man page/
@@ -1069,7 +1112,7 @@ aia --chat -m gpt-4o-mini,gpt-3.5-turbo
1069
1112
  3. envar values over-rides ...
1070
1113
  4. default values
1071
1114
 
1072
- ## [0.3.0] = 2023-11-23
1115
+ ## [0.3.0] - 2023-11-23
1073
1116
 
1074
1117
  - Matching version to [prompt_manager](https://github.com/prompt_manager) This version allows for the user of history in the entry of values to prompt keywords. KW_HISTORY_MAX is set at 5. Changed CLI interaction to use historical selection and editing of prior keyword values.
1075
1118
 
data/README.md CHANGED
@@ -6,7 +6,7 @@
6
6
 
7
7
  <img src="docs/assets/images/aia.png" alt="Robots waiter ready to take your order."><br />
8
8
  **The Prompt is the Code**<br />
9
- <p>Check out the new <a href="https://madbomber.github.io/aia/extending/multi-model/#giving-each-model-a-role">Inline Role Syntax</a> when working with multiple concurrent models.</p>
9
+ <p>New in 2.1: <a href="#coding-agents-as-models-claude-code-codex-copilot">coding agents as models</a>. Point <code>--model</code> at <code>claude_code/</code>, <code>codex/</code> or <code>copilot/</code> and the Claude Code, Codex or GitHub Copilot CLI answers your prompt, alone or as a member of a crew.</p>
10
10
  </div>
11
11
 
12
12
  ---
@@ -33,6 +33,7 @@ For more information on AIA visit these locations:
33
33
  | Prompt management | File-based system with history, variables, and fuzzy search | Limited or none |
34
34
  | Workflow pipelines | Native `--next` / `--pipeline` chaining | Manual chaining required |
35
35
  | Local model support | Ollama and LM Studio out of the box | Varies |
36
+ | Coding agents as models | Claude Code, Codex and Copilot CLIs used like any other model, alone or in a crew | Not available |
36
37
  | Tool integration | RubyLLM tools + MCP server ecosystem | Limited |
37
38
  | Shell & ERB integration | Native — prompts are live Ruby/shell programs | External only |
38
39
  | Agent orchestration | Debate, spawn, verify, decompose, delegate directives | Not available |
@@ -74,15 +75,15 @@ For more information on AIA visit these locations:
74
75
 
75
76
  ```plain
76
77
 
77
- , , AIA v2.0.0 is Online
78
+ , , AIA v2.1.0 is Online
78
79
  (\____/)
79
- (_oo_) Models: gpt-4o-mini
80
- (O) DB: refreshed 2026-03-19 at 10:56
81
- __|||__ \) Libs: ruby_llm v2.1.1, robot_lab v0.5.0, simple_flow v0.4.0,
82
- [/ Tobor \] / trak_flow v0.2.0, typed_bus v0.0.1
83
- / \_______/ \/ Tools: none loaded
84
- / /___\ MCP: (none configured)
85
- (\ /_____\ Crew: @tobor
80
+ (_oo_) Models: claude-sonnet-4-6
81
+ (O) DB: refreshed 2026-11-01 at 22:28
82
+ __|||__ \) Libs: ruby_llm v2.1.1, robot_lab v0.6.0, simple_flow v0.4.0, trak_flow v0.2.0, typed_bus v0.0.1
83
+ [/ Tobor \] / Tools: none loaded
84
+ / \_______/ \/ MCP: (none configured)
85
+ / /___\ Crew: @tobor
86
+ (\ /_____\
86
87
  ::: :::
87
88
  ::: :::
88
89
 
@@ -175,6 +176,7 @@ Implement a schema registry with event-driven synchronization...
175
176
  - [Ollama Integration](#ollama-integration)
176
177
  - [LM Studio Integration](#lm-studio-integration)
177
178
  - [Listing Local Models](#listing-local-models)
179
+ - [Coding Agents as Models (Claude Code, Codex, Copilot)](#coding-agents-as-models-claude-code-codex-copilot)
178
180
  - [Shell Integration](#shell-integration)
179
181
  - [Embedded Ruby (ERB)](#embedded-ruby-erb)
180
182
  - [Prompt Sequences](#prompt-sequences)
@@ -296,7 +298,8 @@ aia --fuzzy
296
298
  | Option | Description | Example |
297
299
  |--------|-------------|---------|
298
300
  | `--chat` | Start interactive chat session | `aia --chat` |
299
- | `--model MODEL` | Specify AI model(s) to use. Supports `MODEL[=ROLE]` syntax | `aia --model gpt-4o-mini,gpt-3.5-turbo` or `aia --model gpt-4o=architect,claude=security` |
301
+ | `--model MODEL` | Specify AI model(s) to use. Supports `MODEL[=ROLE]` syntax. A bare provider name (`claude_code`) means that provider's default model | `aia --model gpt-4o-mini,gpt-3.5-turbo` or `aia --model gpt-4o=architect,claude=security` |
302
+ | `--provider SLUG` | Provider for models named without a `provider/` prefix. Also re-points models from `AIA_MODEL` or the config file at this provider, keeping their names. Scopes `--list-models` | `aia --provider ollama -m llama3.2` |
300
303
  | `--consensus` | Enable consensus mode for multi-model | `aia --consensus` |
301
304
  | `--no-consensus` | Force individual responses | `aia --no-consensus` |
302
305
  | `--role ROLE` | Use a role/system prompt (default for all models) | `aia --role expert` |
@@ -348,6 +351,8 @@ AIA determines configuration settings using this order (highest to lowest priori
348
351
 
349
352
  **Note:** When `-c` / `--config-file` is used, it replaces the user config and environment variables. Configuration resets to bundled defaults, then the specified file is applied, then CLI arguments take precedence.
350
353
 
354
+ **`--provider` sets the provider, not the model.** Models named with `--model` and no `provider/` prefix get it. Models that came from `AIA_MODEL` or a config file get it too, replacing their own prefix but keeping their names, because the command line outranks both. So `--provider claude_code` with `AIA_MODEL=lms/qwen3.8` exported asks Claude Code for a model named `qwen3.8`. To use a provider's default model, name the provider as the model: `--model claude_code` means `claude_code/default`.
355
+
351
356
  ### Configuration Methods
352
357
 
353
358
  **Environment Variables:**
@@ -964,6 +969,46 @@ The `/available_models` directive (alias: `/models`) lists all models known to t
964
969
  - 📡 **Offline**: Works without internet connection
965
970
  - 🔧 **Control**: Full control over model and parameters
966
971
 
972
+ ### Coding Agents as Models (Claude Code, Codex, Copilot)
973
+
974
+ An agent CLI can stand in for a model. With the matching
975
+ `ruby_llm-providers-*` gem installed, the provider prefix sends the prompt to
976
+ the agent's batch mode on stdin and reads its answer back from stdout. The
977
+ agent keeps its own tools, settings and MCP servers and works in the current
978
+ directory, so any aia prompt, pipeline or multi-model run can be driven by an
979
+ agent that reads and edits files.
980
+
981
+ | Prefix | Gem | CLI |
982
+ |---|---|---|
983
+ | `claude_code/` | `ruby_llm-providers-claude_code` | `claude -p` |
984
+ | `codex/` | `ruby_llm-providers-codex` | `codex exec` |
985
+ | `copilot/` | `ruby_llm-providers-copilot` | `copilot` (non-interactive) |
986
+
987
+ ```bash
988
+ gem install ruby_llm-providers-claude_code
989
+
990
+ # The part after the slash is the agent's --model; a bare prefix means its default
991
+ aia --model claude_code/opus refactor_prompt
992
+ aia --model claude_code my_prompt
993
+
994
+ # --provider alone keeps the model name from AIA_MODEL or the config file and
995
+ # hands it to the agent; for the agent's default model use --model as above
996
+ aia --provider claude_code --model sonnet my_prompt
997
+
998
+ # Chat keeps the agent's session alive between turns (--resume)
999
+ aia --chat --model claude_code/sonnet
1000
+
1001
+ # Cross-vendor review: one agent implements, another reviews
1002
+ aia --model claude_code/sonnet=implementer,codex/default=reviewer --consensus review_prompt
1003
+
1004
+ aia --list-models --provider claude_code
1005
+ ```
1006
+
1007
+ Because the agent brings its own tools, aia's `--tools` and MCP servers are
1008
+ not attached to these robots (a one-line notice says so). Permissions, the
1009
+ working directory, timeouts and extra CLI flags are set through the gem's
1010
+ `RubyLLM.config.<provider>_*` options; see each gem's README.
1011
+
967
1012
  ### Shell Integration
968
1013
 
969
1014
  AIA automatically processes shell patterns in prompts:
@@ -1393,18 +1438,18 @@ mcp_servers:
1393
1438
 
1394
1439
  #### Verifying MCP Server Configuration
1395
1440
 
1396
- When MCP servers are configured, AIA displays them in the startup robot:
1441
+ When MCP servers are configured, AIA lists them in the startup robot with each server's tool count in parentheses. Servers that failed to connect are named after `FAILED:`:
1397
1442
 
1398
1443
  ```
1399
- , , AIA v2.0.0 is Online
1444
+ , , AIA v2.1.0 is Online
1400
1445
  (\____/)
1401
- (_oo_) Models: gpt-4o-mini
1402
- (O) DB: refreshed 2026-03-19 at 10:56
1403
- __|||__ \) Libs: ruby_llm v2.1.1, robot_lab v0.5.0, simple_flow v0.4.0,
1404
- [/ Tobor \] / trak_flow v0.2.0, typed_bus v0.0.1
1405
- / \_______/ \/ Tools: 24 tools loaded
1406
- / /___\ MCP: github, htm
1407
- (\ /_____\ Crew: @tobor
1446
+ (_oo_) Models: claude-sonnet-4-6
1447
+ (O) DB: refreshed 2026-11-01 at 22:28
1448
+ __|||__ \) Libs: ruby_llm v2.1.1, robot_lab v0.6.0, simple_flow v0.4.0, trak_flow v0.2.0, typed_bus v0.0.1
1449
+ [/ Tobor \] / Tools: 109 tools loaded
1450
+ / \_______/ \/ MCP: github(42), brew(14), redis(53) | FAILED: htm
1451
+ / /___\ Crew: @tobor
1452
+ (\ /_____\
1408
1453
  ::: :::
1409
1454
  ::: :::
1410
1455
 
@@ -99,8 +99,9 @@ module AIA
99
99
  end
100
100
 
101
101
  opts.on('--provider SLUG',
102
- 'Default provider for models given without a provider/ prefix; ' \
103
- 'also scopes --list-models (any RubyLLM-registered provider)') do |slug|
102
+ 'Provider for models given without a provider/ prefix; also re-points ' \
103
+ 'models from AIA_MODEL or the config file at it, keeping their names. ' \
104
+ 'Scopes --list-models (any RubyLLM-registered provider)') do |slug|
104
105
  options[:provider] = slug
105
106
  end
106
107
 
@@ -162,10 +162,15 @@ defaults:
162
162
  file: STDOUT
163
163
  level: warn
164
164
  flush: true
165
+ # MCP connection logging goes to a file so warnings never interleave
166
+ # with the "Connecting MCP servers" spinner; failures still appear in
167
+ # the startup banner's FAILED list.
165
168
  mcp:
166
- file: STDOUT
169
+ file: ~/.config/aia/logs/mcp.log
167
170
  level: warn
168
171
  flush: true
172
+ # stderr of stdio MCP server processes: a file path, STDERR, or null
173
+ server_stderr: ~/.config/aia/logs/mcp_servers.log
169
174
 
170
175
  # ---------------------------------------------------------------------------
171
176
  # Pipeline/Workflow Configuration
@@ -13,6 +13,10 @@
13
13
 
14
14
  module AIA
15
15
  class ModelSpec
16
+ # The model id used when a bare provider name is given (e.g. `claude_code`):
17
+ # the provider's own default model.
18
+ DEFAULT_MODEL = 'default'
19
+
16
20
  attr_accessor :name, :role, :instance, :internal_id, :provider
17
21
 
18
22
  def initialize(hash = {})
@@ -76,18 +80,23 @@ module AIA
76
80
 
77
81
  private
78
82
 
79
- # Extract a provider prefix from the model name. The prefix is a provider
80
- # exactly when RubyLLM knows it (registered already, or loadable as a
83
+ # Extract a provider from the model name. "<provider>/<model-id>" splits
84
+ # on the slash; a bare "<provider>" with no slash names that provider's
85
+ # DEFAULT_MODEL, which is how agent-CLI providers such as claude_code,
86
+ # codex and copilot are usually given. The prefix is a provider exactly
87
+ # when RubyLLM knows it (registered already, or loadable as a
81
88
  # conventional ruby_llm-providers-<prefix> gem) — no provider names are
82
89
  # hardcoded here. Names whose first segment is not a provider (registry
83
- # ids like "meta-llama/llama-4-maverick") pass through untouched.
90
+ # ids like "meta-llama/llama-4-maverick", plain ids like "gpt-4o") pass
91
+ # through untouched.
84
92
  def extract_provider_from_name!
85
- return unless @name&.include?('/')
93
+ return if @name.nil?
86
94
 
87
95
  prefix, rest = @name.split('/', 2)
88
96
  return unless ProviderRegistry.available?(prefix)
97
+
89
98
  @provider = prefix
90
- @name = rest
99
+ @name = rest || DEFAULT_MODEL
91
100
  end
92
101
  end
93
102
  end