Three new Discipline rules: handle all repos/submodules when given a multi-repo task; diagnose and act, do not ask permission per sub-step; when sandbox-blocked, explain and move on.
Shared agent/tool prompts and skills now live at the top level rather
than duplicated across frontends. Frontend-specific system prompts
remain in their respective dirs (kde/prompts/, dbus/prompts/).
- prompts/, skills/, tools/ removed from monorepo root
- 9p/: gains prompts/skills/tools with 9P filesystem API (unchanged)
- kde/: gains prompts/skills/tools rewritten for D-Bus API
- mkfile: data target no longer installs prompts/skills/tools
- 9p/mkfile: install target now includes data (prompts/skills/tools)
- kde/Makefile: all target now includes data (prompts/skills/tools)
- SYSTEM_PROMPT_QWEN and SYSTEM_PROMPT_DEVSTRAL updated in both
submodules to include the operational world model
- reasoning_think: never call more than once before acting
- Move Planning section from SYSTEM_PROMPT to agent-coding
(generalist doesn't need plan-first mandate)
Move inline JSON prompt strings to prompts/agent-*.md files, referenced
via $OLLIE/x/prime in the new executable prompt array. Move all
prime calls (tools-*, beads) from hooks.agentSpawn into prompt.
Rename greybeard to sysarchitect.
Ollie Agent
Levi Neely
Levi Neely
Levi Neely