Sit an issue#

Open the day’s work#

A plan files the parent and its checklist in one tracker commit, claims the parent, and prints each item with its id.

$ ljos sitting --plan "Ship the codec" -p day -- "measure the files" "pick a default"
day-ab12  Ship the codec
  day-cd34  measure the files
  day-ef56  pick a default
next: day-cd34  measure the files
$ ljos finish day-cd34
ticked day-cd34
next: day-ef56  pick a default
$ ljos ready --mine
next: day-ef56  pick a default

ljos ready --mine names the next unfinished item. ljos finish ITEM ticks that item off. A sitting without --verbose is the short form: what needs action, what the seat remembers, and the claim. A degraded pack is one line, memory degraded: ranked by words. A pack that answers 503, or does not answer, is one line, memory skipped: packsetd busy, and the sitting continues. That search waits at most fifteen seconds when the encoder is cold, and about a second otherwise. Quiet claims are a count of the project file already opened for the ticket.

Start a sitting, and end one#

$ ljos sitting proj-1a2b
== doctor
ok  vissue  ...
ok  seat    alice, holding as alice (from LJOS_SEAT)
== cards
== due
unreviewed  conclusion  3f9c... The lexical default is BM25+.
1 due; 4 scheduled, next at 2026-09-13T09:12:00Z
== island: Ship the fuse change?
1.000   seed    ab8c... 3 weeks ago CombMNZ over RRF for fusing two ballots.
== blockers
none open; the issue is workable
== playbook
sit
A sitting on one issue. Name this recipe at open (`ljos sitting ISSUE --playbook sit` or `ljos playbook ISSUE sit`). The sitting prints this body before recall and holds the name until finish or release.
...
== recall
...
== timeline
timeline of proj-1a2b: Ship the fuse change?
2026-08-30  13 days ago     tracker created
...
== reclaim
no claim quiet longer than 900s
== claim
gen=2

One verb, in the protocol’s order; it stops at the first store that does not answer and claims nothing. --playbook NAME copies a recipe (sit, arena, land, company-panel, overnight, feature) into the playbook section before recall; absent a name, a closed-set token in the title else sit. The name is a tracker playbook: note until finish or release. ljos panel refuses until one is bound. The start_a_sitting prompt gives the same order to a runner that prefers single tools. A claim refused as busy names the issue you still hold; ljos complete finishes it and ljos release hands it back.

$ ljos finish proj-1a2b --lesson "CombMNZ held on turns. RRF lost two points." --outcome ship
completed the session node for proj-1a2b as done
remembered the lesson (id 51778929) as agent-derived; the hook shows it from the next prompt, and `ljos memories review` prunes it
fired the island for "Ship the fuse change?": 8 memories. Those links gained weight under the seat, not under a persona. The next walk of this title follows them.
Learned. 6 trust rows rewritten. ... persona anchors moved. The outcome is kept: once 6 issues have one, ljos consensus discounts voters who err together. ... No stated probability, so there is no Brier score. A hard vote is not a claim of certainty.
the ticket proj-1a2b stays open and nobody holds it; a person closes it when the work is accepted

Memories, a no vote, and a child#

A seat remembers with no sitting open. --as NAME stores the lesson in the set persona-NAME. Each persona’s brief then carries that persona’s lesson, and the lists are not the same.

$ ljos remember --as release-engineer "Cut a release only after the changelog names the search index."
21e499e4b8487a3abc7ff7479e9da102    lesson  due 2026-11-02T14:29:43.738Z    Cut a release only after the changelog names the search index.
standing: the hook shows it from the next prompt it bears on
$ ljos remember --as docs-reader "A newcomer looks for the search index in the notes."
bac3e6e51514703c76672d163cc38cde    lesson  due 2026-11-02T14:29:43.770Z    A newcomer looks for the search index in the notes.
standing: the hook shows it from the next prompt it bears on
$ ljos brief release-engineer demo-0gmm
What you remembered yourself:
- [lesson, agent-derived, today] Cut a release only after the changelog names the search index.
$ ljos brief docs-reader demo-0gmm
What you remembered yourself:
- [lesson, agent-derived, today] A newcomer looks for the search index in the notes.

A no vote has to name an open child. ljos file --parent links it in the tracker. The parent stays open while that child is open. After the child is closed, withdraw the no ballot and vote again, then ljos consensus.

$ ljos vote demo-0gmm --against "the notes omit the index" --used none --as perf-reviewer
Error: vote: a no ballot on demo-0gmm has to name a child ticket. `ljos file --parent demo-0gmm "TITLE"` files one, then `ljos vote demo-0gmm --against REASON --child ID` links it. Nothing was cast
$ ljos file "Name the search index in the notes" --parent demo-0gmm
demo-1zcv  TODO  [#C]  Name the search index in the notes
$ ljos vote demo-0gmm --against "the notes omit the index" --used none --as perf-reviewer --child demo-1zcv
recorded on the tracker: perf-reviewer for no on demo-0gmm
no ballot linked to child demo-1zcv. demo-0gmm stays open while demo-1zcv is open. After `ljos finish demo-1zcv --close`, replace the ballot: `ljos vote demo-0gmm --withdraw --as perf-reviewer` then `ljos vote demo-0gmm --for OPTION --used none --as perf-reviewer`, then `ljos consensus demo-0gmm`.

Hold a done claim beside a red test#

Stop also blocks one turn, with no model in the loop, when the conversation has no issue and the turn used tools without touching the seat. The agent is told to file an issue and sit. The next stop of that same turn goes through.

With Jev on, the Stop and SubagentStop hooks audit the turn once before the agent stops. The audit reads four parts of the turn, bounded to a few thousand tokens:

the person’s last request
the shell commands the turn ran since that request
the latest test output, or the last output when no test ran
the final message from the agent

Jev answers three questions about it:

question	blocks the stop at
`claims_complete`	0.9 or above, with `tests_green` 0.1 or under and a test run in the turn
`tests_green`	(with the above)
`deferral`	0.9 or above

Whether a test ran is decided in code, from the command line. A held agent gets the reason and one more round; its second attempt to stop is never audited, so it cannot loop. A final message under 80 characters is not sent. The cuts sit in jev.rs beside the questions. A stop costs about 3,000 input tokens, or $0.00013.