dingger.git/crates/dinggerd/src/hooks.rs, branch master

show each pi agent's context use: a meter in the panel and search, text in the menu

2026-11-03T11:05:57+00:00

show what pi agents cost today, this week and this month, counting forked replies once

2026-11-01T08:22:01+00:00

mark forgotten drafts in pi agents' input boxes, and keep them across a crash or restore

2026-10-30T16:59:43+00:00

reload pi agents on an old extension: versioned extension, control socket, menu item and reload-agents

2026-10-30T15:27:21+00:00

mark as unread: make an idle agent done again, to come back to it

2026-10-27T12:12:28+00:00

like "mark as unread" in mail: an agent you have seen (idle) becomes done again, last among done agents of its importance, with no sound or notification. it stays done until you come back to it: leave its tmux window and return, or jump to it. marked while you look at it, it holds until you look away (`held`); a new turn clears it. only idle agents can be marked. - daemon: `pane.mark_unread` on the hook socket; the watcher applies it after the tick's observations (mark_unread, advance's hold rule). - `dingger unread [--pane]`: refuses done, working, blocked, and panes without an agent; returns once the agent shows as done. - pi: /unread (refused while the agent works or is blocked). - panel: right-click an idle agent in the full list, "Mark as unread". - search: ⌘U or right-click marks the selected idle agent; the footer offers ⌘U when it applies; search stays open. tests: unit (hold, jump, new turn, idle only), e2e step 22 (queue position, hold, look away and back, jump, refusals, no sound, menu, the extension), pi-live in real pi 1.4 (a mark, not a turn: an unknown command would also end done), swift selftest (menu only for idle, footer).

hooks: answer each message, so the extension sends a report once

2026-10-26T14:40:50+00:00

the pi extension (herdr's protocol) waits for a reply line and sends a report again after 0.5 s without one; dingger never replied. every report arrived twice and took about 2 s, delaying the next one (the live log: 196 of 400 reports repeated the previous one, 0.49 s later). the daemon now answers each line with {"ok":true}; senders that do not read replies (dingger poke, importance, the claude hook) are unaffected. e2e step 21: one report, done in 0.38 s; without replies, two, 2.08 s.

importance: mark an agent high, normal, or low

2026-10-26T14:32:18+00:00

within a status, more important agents are listed first, and so jumped to first; status stays first. a high agent shows a bold ! before its name, a low agent's name is lightly dimmed with "low" on its second line, and it makes no sound (it still notifies and waits in line). in pi, /importance high|normal|low (completes the levels; none once one is typed out, so Enter runs it). the extension saves it in the session as a custom entry (dingger-importance), so /reload, restarts, /resume and restore bring it back; /new starts at normal. it shows a level other than normal in pi's footer, emits dingger:importance { level } for other extensions, and sends it with every report. for other agents, `dingger importance [level] [--pane]`; it refuses a pi that reports itself, whose saved level would win. the daemon stays the one writer of pane state: it writes @dingger_importance (unset for normal), which goes with the agent and survives a daemon restart. `dingger status` marks high and low. tests: e2e step 21 (order, the option, no sound for low, the menu bar, the command, the extension through a harness that now matches pi's extension API), a DINGGER_SOUND seam so sounds are observable, and a status-line check that read the counts once (flaky) now polls. pi-live: /importance in real pi 1.4, its footer and session entry, and the level after a restore; every pi there now uses the test's agent directory (a restored pi read the user's). negative controls fail each part.

restore Claude Code agents in their own sessions

2026-10-23T12:08:53+00:00

`dingger install claude` adds a SessionStart hook to Claude's settings (every other key and hook kept, in order; the original backed up; written through a symlink; installing again changes nothing). The hook is `dingger hook claude`: it reports the session's transcript for its tmux pane, prints nothing (Claude adds SessionStart output to the conversation) and always exits 0. Claude's state still comes from its screen. restore types `claude --resume ` in the directory the session ran in (elsewhere, Claude moves the conversation there). it resumes an agent only when its session file exists: a Claude session that never got a message has no transcript, and `--resume` fails ("No conversation found"). the snapshot now keeps an agent's command line too, so such a pane gets it retyped instead of nothing. - the hook receiver takes a report without a state: only the session. - a report that arrives after the watcher listed the panes is not dropped as "agent gone": Claude reports once, as it starts. - serde_json preserve_order (indexmap is already a dependency, via toml). e2e step 18: install against a settings file like a real one, a fake claude running the exact installed hook command, resume by id, retype without a transcript, the resumed agent reporting again. a server that restore starts now gets the test's plain shells: it got the user's login shell, whose profile put the real claude first, which started (stuck at its trust dialog) during development. every test restore now checks that no real pi or claude runs in a test pane.

restore, step 2: resume agents, retype command lines

2026-10-23T10:20:11+00:00

- the pi extension reports its session file; the daemon keeps it in the pane option @agent_session, and snapshots carry it. restore runs `pi --session ` in the rebuilt pane, one agent every 0.5s, then reports which agents run and which were interrupted mid-turn. - other panes: the command line running in them is typed at the new prompt but not run. exactly as typed with the zsh hook (dingger install zsh: preexec/precmd keep @dingger_cmd), else the foreground program's exact argv from the kernel (KERN_PROCARGS2), shell-quoted. - "at a prompt" is the pane's own shell owning the terminal's foreground process group, not a name check: `sh -c ...` is a running program. - an agent or command that exits keeps its place in snapshots for 60s, so a shutdown that stops programs before tmux still restores them. - typing waits for each new prompt (text typed earlier echoes twice). e2e step 16 covers resume with a session path containing a space, the interrupted report, an agent that exits just before the server dies, a quoted command retyped not run, and the hook's text; dropping the grace period, judging prompts by name, running retyped commands, or not storing the session reference each fail it. the test's fake pi is guarded so a real pi never starts.

restore, step 1: layout snapshots and dingger restore

2026-10-23T10:01:52+00:00

the daemon saves the tmux layout when it changes: sessions, windows (index, name, automatic-rename, layout, size, zoom, active), and panes (directory, active, agent). a new history file for layout changes, a rewrite for details, the last 50 kept, nothing saved while tmux is down. when a different server runs (or none), the old server's last snapshot is pinned as the restore candidate before the new server's snapshots start. dingger restore rebuilds missing sessions by tmux id: panes split from the newest so they keep their order, the saved layout applied at the saved size, then the size released so windows follow clients again. a bare session of the same name (a terminal's `new-session -As 0`) is replaced, its clients switched over first; used sessions are kept. --dry-run, --list, --from, --session. e2e step 16 kills a server with a client attached like ghostty's, restarts it, restores, and compares every window and pane; splitting the wrong pane, not replacing bare sessions, not pinning, or keeping the saved size each fail it. e2e keeps its state in its temp dir.