Commit graph

12 commits

Author SHA1 Message Date
Alessandro
db8bc111ca prompts: strict json guidance; concise responses
Trying to steer the model into treating the final curly brace of a JSON obj as EOS token. Also defaulting to concise responses, expanding only when needed.
2026-04-03 14:13:37 +02:00
Alessandro
86dca86f6f prompts: restore legacy, plugins, agent0 profile
Restore main.communication, main.solving, main.tips and tool.response, which made the model dumber. For some reasons the drawback was more visible with frontier LLMs.

restore builtin plugins and agent0 profile
2026-04-03 06:49:25 +02:00
Alessandro
74efea68cd prompts: override role with specifics md file 2026-04-01 18:47:36 +02:00
Alessandro
54362bf8ee prompts: adopt compact defaults and rebalance guidance
Promote the a0_small prompt work into the shipped default stack.

Token count for default prompt, accounting for tools, plugins, secrets placeholders, inactive project, etc: 2996 tokens.

- migrate compact prompt fragments into core prompts, plugin-owned prompts, and the agent0 overlay
- remove the obsolete a0_small profile and rename the tool-call knowledge reference to a generic framework file
- keep a small set of high-value JSON examples while preserving key guardrails for tool names, tool args, and method-style calls
- bake balanced conciseness into default response behavior without changing the initial greeting
- drop the call_sub prompt-side loader and keep call_sub guidance self-contained
- restore essential secrets/security guidance, preserve inactive-project messaging, and rebalance the A2A prompt
- add a regression test that verifies the assembled default agent0 prompt stays under the token budget and retains critical guardrails
2026-03-30 21:44:20 +02:00
Alessandro
a2bd044a9d a0_small prompt set prototype
Created a new prompt set for small models.

A0 now starts with 2.5k tokens usage, not 10k.

improve small model tool policy

prompts: knowledge with tool call examples

- Adds knowledge examples for tool-call shapes and memory usage (workaround to not have examples in tool prompts)
- Adds a small prompt hint to consult memory examples when tool args are unclear

stricter small-model guardrails

Canonical JSON example, exact tool-name matching, method suffix enforcement, British spelling for behaviour_adjustment, boolean guidance for call_subordinate.reset, and the “don’t improvise tools after an error” rule are the new additions here.

Trying to make
2026-03-30 07:11:34 +02:00
frdel
d02dda3667 BIG PYTHON REFACTOR
Python scripts moved out of python/ folder to root to be unified with plugins

+ frontend extension around api calls
2026-03-05 17:28:11 +01:00
frdel
6e2a442222 Convert agent and plugin metadata/configs to YAML
Replace JSON-based agent and plugin metadata/config files with YAML equivalents and update code paths to read/write YAML. Added python/helpers/yaml for YAML load/dump and files.read_file_yaml helper; subagents, plugins, and files helpers now prefer agent.yaml/plugin.yaml and default_config.yaml. Migration now includes convert_agents_json_yaml and startup_migration is invoked from initialize to perform conversions. Plugin toggle/config handling was refactored (META_FILE_NAME/CONFIG_DEFAULT_FILE_NAME changes, new determined_toggle_from_paths helper) and memory plugin defaults and extensions were updated to use per-agent plugin config lookups with guards for missing configs.
2026-02-23 19:36:45 +01:00
frdel
fd06f51ee0 polish skills import, bugfixes 2026-02-06 13:41:56 +01:00
frdel
3adec910c0 image response prompt, image view scroll 2026-02-02 20:19:35 +01:00
frdel
11d1cb0e64 subagents preparation 2025-11-26 09:07:30 +01:00
frdel
0aacdf98bc v0.9.5-pre cleanup, polishing, bugfixing 2025-08-25 09:59:44 +02:00
frdel
49c06193c2 agents rework, memory improvements and params 2025-07-25 21:47:19 +02:00