All issues

Meta: revamp agmsg.cc — coordination thread (raise your hand here)
#213 · fujibee opened on Jun 24, 2026
4

peek's confirmed-gone check reads herdr's error by substring, not structure — a renamed error code silently becomes 'unknown' instead of 'gone'

#1169

· fujibee opened

on Sep 12, 2026

herdr: a socket path containing ':' cannot round-trip through the pane locator

#1166

· fujibee opened

on Sep 12, 2026

The Windows bridge-launcher test fails intermittently and blocks landing, while the issue it names is closed

#1161

· fujibee opened

on Sep 11, 2026

peek discards the terminal error and reports every read failure as "the pane may no longer exist"

#1158

· fujibee opened

on Sep 11, 2026

Two canonical facts: the seat's agmsg id and its current pane. Everything else is derived, and the seat writes it

#1152

· fujibee opened

on Sep 11, 2026

A plain member is launched with no placement record, so nothing can reach, repair or tear it down

#1149

· fujibee opened

on Sep 10, 2026

Two seats can claim one pane and nothing reports it; --fix repairs each seat and leaves the collision

#1144

· fujibee opened

on Sep 10, 2026

A placement record written by the action hook has no project or type, and arrange refuses it

#1137

· fujibee opened

on Sep 10, 2026

A seat can find its own pane by emitting a token and seeing where it appears

#1124

· fujibee opened

on Sep 10, 2026

codex sessions keep their generated name: --fix skips the rename it can actually perform

#1118

· fujibee opened

on Sep 9, 2026

Every codex seat resolves its pane to the one the app-server was launched from

#1112

· fujibee opened

on Sep 9, 2026

team --fix types into sessions, and its name does not say so

#1110

· fujibee opened

on Sep 9, 2026