buzz-acp: tell the agent to start fresh on the session/new that follows !rotate - #6103
Open
jhgaylor wants to merge 4 commits into
Open
buzz-acp: tell the agent to start fresh on the session/new that follows !rotate#6103jhgaylor wants to merge 4 commits into
jhgaylor wants to merge 4 commits into
Conversation
This was referenced Aug 17, 2026
11 tasks
|
Same conflict as #6104 — this branch also collides with (Context: auditing a cluster of concurrent |
…nel on session/new A harness restart — a desktop relaunch, a config-change restart, or a hosted harness moved between nodes on a deploy — emptied the in-memory channel → session map, and the next mention in every channel got a fresh session/new. For an agent whose session *is* its workspace (a sandbox per session, as with the `fountain acp` gateway) that discarded the channel's memory and files each time; the base prompt tells the agent to "resume silently after a session restart", which assumes the opposite. Two changes: * `SessionStore`: a small JSON file (`<state-dir>/sessions-<pubkey>.json`, atomic writes) remembering channel → session id. Before opening a new session for a channel with no live one, the pool `session/load`s the remembered id when the agent advertised `loadSession` at initialize; any failure forgets the id and falls back to `session/new`, so a resume that cannot happen never costs the turn. Leaving a channel and `!rotate` forget the entry. `--state-dir` / `BUZZ_ACP_STATE_DIR` (default `<cwd>/.buzz-acp`), `--no-resume-sessions` / `BUZZ_ACP_NO_RESUME_SESSIONS` to opt out. * `session/new` now carries `_meta.channelId` and `_meta.channelType` (`AcpClient::session_new_with_origin`; `session_new_full` is unchanged and delegates). The prompt names the channel only as text; an agent or gateway that wants to key sessions by channel — to resume the same conversation after this harness restarts even when its own state directory is not persistent — had nothing machine-readable to key on. Absent for heartbeat sessions. Tests: store round-trip/malformed/disabled; client sends channelId/type and merges into existing _meta; loadSession capability recorded; session/load request shape and tolerance of the pre-response replay; error surfaces; config defaults and opt-out. Not covered here: a scripted end-to-end through the pool's get-or-create path (the pool has no scripted-agent test harness today). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Signed-off-by: Jake Gaylor <jhgaylor@gmail.com>
…container volume note Reviewers asked for the README configuration table to list the two settings introduced with session resumption, and for an operator note that the default <cwd>/.buzz-acp state dir is ephemeral in containers and must be pointed at mounted persistent storage. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Signed-off-by: Jake Gaylor <jhgaylor@gmail.com>
Desktop and mobile insert a mention into the message body as literal text
("@Fountain Maintainer !rotate") alongside the p tag, but
is_owner_control_command required content.trim() to equal the command
exactly. So a mentioned command fell through to the agent as a prompt, and
a bare command had no p tag and was dropped — the commands were unreachable
from every product surface (already noted in
docs/welcome-kickoff-silent-failures.md §5).
Match the command when it is the whole content, or when it is the last or
first token with only @name / nostr: mention text on the other side. Content
that continues past the command is still forwarded as an ordinary message.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Jake Gaylor <jhgaylor@gmail.com>
…ws !rotate !rotate drops the harness's ACP session for a channel, and the next turn sends session/new. But session/new carries _meta.channelId, and an agent or gateway that keys its own state by channel — fountain acp's channel-bound conversations — hands the same conversation straight back. On such an agent !rotate was a no-op: observed on the hosted Fountain Maintainer, where a !rotate logged as consumed and the next turn resumed the same conversation. The harness now remembers the rotate (SessionStore::request_fresh) and the next session/new for that channel carries _meta.freshSession: true, which the agent must honour by not resuming. The member is absent otherwise. Also: an in-flight !rotate never removed the channel from the on-disk session store (only the idle path did), so with a state dir the post-rotate turn could session/load the rotated session. Both paths now forget it. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Signed-off-by: Jake Gaylor <jhgaylor@gmail.com>
jhgaylor
force-pushed
the
buzz-acp/rotate-requests-fresh-session
branch
from
August 18, 2026 02:47
6daa3f5 to
20f34e6
Compare
Author
|
Thanks @mfethe1 — rebased the whole stack (#6088 → #6101 → #6103 → #6104) onto current |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
!rotatedrops the harness's ACP session for a channel, and the next turn sendssession/new. Butsession/newcarries_meta.channelId(#6088), and an agent or gateway that keys its own state by channel —fountain acp's channel-bound conversations — hands the same conversation straight back. On such an agent!rotateis a no-op.Observed on a hosted harness:
!rotatelogged as consumed (invalidated=1), and the next turn resumed the same conversation.Change
SessionStore::request_fresh(channel)/take_fresh(channel)— remember that the owner rotated the channel (in-memory; independent of whether the store is persisted).!rotatepaths (idle and in-flight) callrequest_freshandsession_store.removeup front.session/newfor that channel carries_meta.freshSession: true(SessionOrigin.fresh); the member is absent otherwise. Consumed on that one call.Also fixes a latent bug from #6088: an in-flight
!rotatenever removed the channel from the on-disk session store (only the idle path did), so withBUZZ_ACP_STATE_DIRset, the post-rotate turn couldsession/loadthe rotated session.The consuming side for Fountain is BinaryBourbon/fountain#788.
Test plan
session_new_with_origin_sends_fresh_session_after_rotate(new),session_new_with_origin_sends_channel_id_and_type_in_metaasserts the member is absent when not rotatedsession_store::fresh_request_is_consumed_once_and_works_when_disabled(new)cargo test -p buzz-acp --lib— 791 passed; clippy + fmt clean@agent !rotateon Fountain → next mention lands in a new conversation🤖 Generated with Claude Code