✨ Add `xmd prompt`: turn a request into an approved executable Plan (#260) - #657

Merged

taras merged 15 commits into

mainfrom

agent/issue-260-prompt-command

Aug 30, 2026

Commits on Aug 30, 2026

✨ Add xmd prompt, the authorship command (#260)
taras
committed
de53362
🐛 Settle one agent stack, refuse -e, and report a run's failure (#260)
taras
committed
d6ca47d
✨ Ship the plan program as a packaged Markdown document (#260)
taras
committed
ddd2999
🐛 Make plan approval and exhaustion mutually exclusive (#260)
taras
committed
a22c7b1
✨ Turn a request into a Plan with a packaged command document (#260)
taras
committed
e06e783
🐛 Give each prompt session its own directory and a reachable exhaustion (#260)
taras
committed
9ee0c40
🐛 Own the profile root in tests, and give a default session's directory back (#260)
taras
committed
6348b73
✨ Capture the problems with <Json as>, and claim the directory before making it (#260)
taras
committed
1e44987
✨ Make the approved Plan the result, and the workflow readable end to end (#260)
taras
committed
7f2a10f
🐛 State the whole Plan requirement in every turn that asks for one (#260)
taras
committed
869a001
🐛 Use the approved apostrophe, and refuse --save by name (#260)
taras
committed
0a927b4
🤖 Say that the second root execution is optional (#260)
taras
committed
142272d
🐛 Refuse a valued --run instead of reading it as the switch (#260)
taras
committed
944c734
🐛 Assert the session and output-file tripwires on the --run= refusal (#260)
taras
committed
1e782a8
🐛 Pin the valued --run refusal at the parser boundary it crossed (#260)
taras
committed
7d0889f