- BaseAgentAdapter: id_lifecycle (pre_assigned: claude, grok / post_discovered)
and capture_active_session_id - exactly one verified candidate newer than the
session that no other row records, else a reason (no-candidate, ambiguous, ...).
- agy: verify_artifact (discover) rejects an id last_conversations.json maps to
another workspace and always excludes ids other rows record (a cache hit no
longer bypasses it). capture uses the cache only as a proposal: the
conversation's presence lock or DB must be held by the row's pane process or
a descendant (lsof/ps); anything unprovable is not pinned.
- lib_py/agents/autopin.py: handle_completed, run by the side that verified the
job's completed - the delegate-job wrapper (direct), create --onboard and the
loop's wait_for_job (recovery, as a subprocess: G9). Local herdr jobs only;
one (job, dispatch token, attempt) is handled once (job-scoped flock + marker
in the job dir); writes go through atomic_dump_yaml; a pin records
session_id_source=discovered / session_id_verified=true; a failure leaves the
reason on the row and never changes the job's result.
- reconcile.sh drift C for agy, hermes, opencode and codex calls the same
capture (the agy *.db glob is gone); _pin_and_verify_resume uses pin_row.
- Tests: tests/test_issue40_auto_pin.py; issue64 :407 now expects False; the
static reconcile-source checks follow the move to the adapters; the two tier2
drift-C component tests create with --no-onboard (onboarding now pins).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
opencode 2.x needs no login: a fresh HOME runs a free default model and a
local provider has no key, while 2.x keeps no auth.json. auth_ok is now
"opencode on PATH or under <HOME>/.opencode/bin". `python -m lib_py.agents
auth-ok <agent>` exits with an adapter's verdict, and create's opencode
branch uses it (same 7 lines) instead of its own auth.json/key test and
`opencode --version`, which wrote a line to the real opencode log.
test_a4's opencode block now pins the binary rule (decision D1, required
by the brief's item 3 and by both reviewers). Other agents are unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Measured on opencode 2.0.32 (macOS):
- the TUI rejects --agent ("Unrecognized flag"), so create and resume could
not start opencode at all; the default agent is build. MAM now spawns
`opencode --auto --standalone` (resume adds --session <id>): the server
runs as the TUI's child and ends with it, and the per-user background
service is neither used nor started;
- sessions live in session_v2 (a fresh 2.x database has no `session` table;
a migrated one keeps the 1.x tables but no longer writes them). discover,
verify_artifact, reconcile drift C and status read session_v2, and
discover / drift C skip child (subagent) sessions;
- `db path` is gone (it prints the help and exits 0): the db comes from
the `db` line of `opencode debug paths`, else the default path;
- `opencode session delete` removes a session with its child sessions from
the 2.x tables (and leaves the 1.x ones): purge runs it with --standalone
for a session in session_v2, still deletes the 1.x message/part/session
rows, and residual also reports a session_v2 row left by a failed delete.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Review loop (--review-only, --all-reviewer, focused on this Mac:
macOS, /usr/bin/python3 3.9.6, bash 3.2): round 1 (@6f4f40a) agy and
grok PASS, claude and codex NOT PASS (a non-uuid id `memory` purged the
auto-memory dir, a linked dir below a root was followed, a NaN
startedAt counted as dead), sent as a PR comment and fixed by the
author in eleven commits; round 2 (@2a8d738) confirmed all three fixed
but 4/4 NOT PASS on one test holder (`sh -c "sleep 120" <uuid>` loses
the uuid on macOS /bin/sh), fixed in de6f3fc; round 3 (@de6f3fc) 4/4
PASS.
Full suite on macOS / Python 3.9.6 @de6f3fc: 1934 passed, 56 skipped,
1 error (the known #55 false positive: agy conversations of another
workspace). The real ~/.claude, /private/tmp/claude-<uid> and herdr
panes unchanged.
Also names lib_py.delegation.logroot in eager.CORE_MODULES (the v4.9.4
review's non-blocking note).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Review loop (--review-only, --all-reviewer, focused on this Mac:
macOS, /usr/bin/python3 3.9.6, bash 3.2): round 1 (@63d8164) agy and
grok PASS, claude and codex NOT PASS (F1 paho via the /usr/bin/python3
fallback, F2 logs default path, R1 disk-fallback redispatch
interleaving, R2 symlinked registry), sent as a PR comment and fixed by
the author in four commits; round 2 (@a7cb90e) 4/4 PASS.
Full suite on macOS / Python 3.9.6 @a7cb90e: 1867 passed, 56 skipped,
1 error (the known #55 false positive: WAL/SHM of an agy conversation
running in another workspace). Real herdr panes and ~/.claude/projects
unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
from_env() fills delegate_job_logs_dir with the delegate_job_logs next to
the resolved mam_jobs_dir unless DELEGATE_JOB_LOGS_DIR is set, so a session
created with only MAM_JOBS_DIR moved no longer records the workspace's old
log root. worktree_prepare.sh passes DELEGATE_JOB_LOGS_DIR on only when the
caller set it. lib.sh and the delegate-job wrapper no longer carry a bash
copy of the default: delegate_publish_cmd, delegate_started_recorded and
_logs_dir_abs resolve it with get_logs_dir(<registry>), the wrapper fails
closed when it cannot, exports the resolved root to its children, prints
it as submit's last line, and logs takes --registry-dir.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Review loop (--review-only, --all-reviewer): round 1 (@3a30d7d) 4/4 PASS;
each reviewer reproduced W1 failing on the pre-fix tree. Seven #50/#55 and
leak modules (112 tests) from a clean worktree left no .mam in the
checkout.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Review loop (--review-only, --all-reviewer): round 1 (@325b056) 4/4 PASS,
round 2 (@db7f4ca, HERDR_SESSION) 4/4 PASS. Full suite at 325b056 on
macOS / Python 3.9 / bash 3.2 / herdr 0.9.3 from an agent pane:
1845 passed / 0 failed / 56 skipped (18 test_deploy_freshness skips: no
nats-docker submodule in the fresh worktree); the panes of the real
servers (default, multi-agent-mux) and ~/.claude/projects were unchanged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Merge review on macOS (--review-only loops, --all-reviewer): round 1 at
320aa9f 4/4 NOT PASS, round 2 codex F1 (a claimed dry-run job holds its
session), round 3 4/4 PASS (discarded: the worktree changed mid-review),
round 4 at 21f0c12 claude/agy/grok PASS with codex interrupted by a stream
disconnect, round 5 codex PASS. Full suite at 21f0c12 on macOS / Python
3.9 / bash 3.2 / herdr 0.9.3: 1860 passed / 0 failed / 38 skipped; no new
entries in ~/.claude/projects.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
- admission: a `submit --dry-run` record is marked dry_run and is not an
open job of its session (review 1e25cf2f F1)
- create: the monitor-row adoption requires the row's server to be the one
this create spawned on (review 1e25cf2f F2)
- pane lock: retry only on contention (BlockingIOError); the docs say the
flock lasts while any fd-9 holder (the claim shell and the shim) lives
(review 66680d9d)
- tests on macOS: real-herdr box HOMEs under /tmp (sun_path is 104 bytes),
resolved paths, the fake claude reads non-canonical input (MAX_CANON 1024)
and the pane is read unwrapped; lock holders exec their sleep so killing
them frees the lock; the adoption tests name the server
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Three `create --onboard` at once into one folder (the control root, so no
control-plane prefix reaches the pane and the shim reuses an idle pane):
the shim's new-session chooses a pane - an idle one in a workspace on the
same cwd, a split, a fresh workspace's root pane - and then claims it with
`agent start`; two shims could choose the pane one of them had just made,
and the later start failed with agent_pane_busy (one create in three).
mam_herdr now runs every new-session call holding one lock per server: a
flock on fd 9, opened for the block and taken by lib_py.pane_lock, which
also bounds the wait (MAM_PANE_CLAIM_WAIT_S, default 120). It is released
with the block or with the process, however it ends. The shim heredoc is
unchanged (S11). Other servers' creates are not held.
Serialising the claims exposed a second race: a create's end-of-run
`reconcile.sh --once` adopted the pane of another create that had not yet
written its row, and that create refused its own name as already running
(exit 4). Such a row - registered by the monitor, no job, for a pane
created after this create began - is now replaced, not refused.
Tests: real herdr - shim-level and create-level three-way races into one
folder (the real herdr's pane-creating replies held back so the window is
hit on every run; both fail on the base with the orchestrator's error),
the different-folder test kept; pane_lock unit tests (per-server file,
timeout, kill -9 holder, handover, knob); the adoption rule and its three
refusals; shim contract h26 (held lock: no pane calls). The herdr box gets
a ~/.zshrc: without it every first `agent start` typed into zsh's first-run
menu and timed out at 30 s.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The handshake only said the agent saw the prompt (#23). create_session.sh
now waits, after `started`, for the onboarding job to reach a terminal
status and makes that its own result, so the orchestrator learns it from
create's exit (MULTI_AGENT_RULES §6) instead of guessing from herdr's
agent_status:
- lib_py/delegation/wrapper.py: `wait-terminal` (completed|error|cancelled,
`timeout` after --timeout, `missing` when the record is gone; stops with
its caller via --watch-pid) and `onboard-exit` (completed 0, error or
cancelled 44, timeout or missing 45). bash branches on the printed answer.
- lib.sh: delegate_wait_terminal / delegate_onboard_exit, thin calls.
- create_session.sh: `onboarding completed: job <id> (<n>s)` on stdout, or
`last_visible_status: onboarding-<status>` on the row, the reason on stderr
and `exit 44` / `exit 45` after `trap - EXIT` - the session is kept, as for
a missing handshake. The "no started" path (error, exit 0 before) now ends
with 44. MAM_ONBOARD_WAIT_S (default 600, floor 1) bounds the wait.
- worktree_prepare.sh: 44/45 pass through with "session and worktree kept".
- create SKILL.md exit table (44, 45) and the knob; the #42 inventory follows.
- tests/conftest.py: --onboard is the default, so ~60 sandbox create tests
reach the new wait. The mock herdr now acts as the agent for create's
onboarding prompt - it runs the publish commands the prompt names through
the real publish_event.py - and MAM_ONBOARD_WAIT_S=3 caps a test that turns
that off (MAM_MOCK_ONBOARD_HANDSHAKE=0 -> 44, `started` -> 45). The
delegate-job wrapper's prompt has another shape and is never answered.
- tests/test_issue50_onboard.py: the decision tables, create in the sandbox
(0 / 44 / 45, session kept, row note, trap window), the knob's integer
rule, worktree_prepare's pass-through.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Merge-workspace review loop (--all-reviewer): round 1 (@a93c9bf) 4/4
PASS. Full suite at a93c9bf on macOS / Python 3.9 / herdr 0.9.3 with
WORKSPACE_ROOT set to an empty sentinel: 1752 passed / 0 failed / 38
skipped; the sentinel stayed empty, the review worktree had no .mam, and
the main checkout's shim was not touched.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Merge-workspace review loop (--all-reviewer): round 1 (@d198dd5) claude
agy PASS, grok codex NOT PASS (the same inherited WORKSPACE_ROOT in eleven
more modules; test_orc_onboard's import-time env); fixed in 8b47cc7.
Round 2 claude grok NOT PASS (test_o2_race_free_lock, test_o1_rebuttal
missed), agy codex PASS; fixed in ac48cb7. Round 3 (@ac48cb7) 4/4 PASS.
Full suite at ac48cb7 on macOS / Python 3.9 / herdr 0.9.3 with
WORKSPACE_ROOT set to an empty sentinel: 1751 passed / 0 failed / 38
skipped; the sentinel stayed empty and the main checkout's shim was not
touched.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Merge-workspace review loop (--all-reviewer): round 1 (@8241bd5) claude
grok PASS, agy NOT PASS (two static wrapper-literal tests did not follow
_decided/_exit_for), codex NOT PASS (R1: a yes/no decision that could not
run read as "no"); fixed in 541c3fe and 65b78b6. Round 2 (@65b78b6) 4/4
PASS. Full suite at 65b78b6 on macOS / Python 3.9 / herdr 0.9.3: 1730
passed / 0 failed / 56 skipped; 18 of the skips were the uninitialized
nats-docker submodule in the fresh worktree, and with it
tests/test_deploy_freshness.py is 31 passed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Merge-workspace review: round 1 (@518beda) claude agy grok PASS, codex
NOT PASS (finalize wait measured on the wall clock); fixed in 4c86107.
Round 2 delta (@4c86107) claude agy grok PASS; codex (24b21b9c, after
817d1650 closed deadline_timeout with no report) found R1 closed and no
new defect, but withheld PASS only because its sandbox could not bind
the suite's loopback broker. The same two-file suite run here at
4c86107: 278 passed. Full suite at 518beda on macOS / Python 3.9:
1711 passed / 1 failed / 38 skipped; the failure
(test_issue45_restore, herdr 0.9.3 cwd restore) also fails on main.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
tests/test_issue45_herdr_config.py imported tomllib at module level. On
Python 3.9 - macOS's /usr/bin/python3, which the venvs and the bootstrap's
python3 use here - the import failed during collection and pytest ran no
test of the whole suite.
The import is optional now, and the 29 tests that parse a result with
tomllib or expect the user's settings kept are marked needs_tomllib: on
Python < 3.11 herdr_config deliberately writes the key-only config (restore
still off, the other settings dropped, a warning on stderr), which
test_without_tomllib_restore_is_still_off covers on every Python.
Python 3.9: 22 passed / 29 skipped; Python 3.11: 51 passed.
create SKILL.md (same line, so the #42 inventory is unchanged) and the
4.8.5 entry state that building the server config needs tomllib (Python
3.11+) in the python3 on PATH, and what an older python3 loses. Checked
with Python 3.9 in a throwaway HOME: the warning reaches stderr (the
bootstrap captures only stdout) and the user's file is untouched.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Issue #45 A+B: MAM's isolated herdr servers start with restore-on-restart off,
and resume closes the session's own restored empty shell. Not reviewed yet.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The .claude/skills check accepts both local layouts and fails on a dangling
link. Review: round 1 (@e1ed270) claude agy grok PASS, codex NOT PASS (a
dangling top-level link was skipped); fixed in 150b493; round 2 4/4 PASS.
Reports and the round-2 identity sidecars go to .agents/reports/<session>/.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Issue #42 stages 1-2: the SKILL.md code-block inventory as test data and
the removal of blocks the entry scripts already do. Merge-workspace review:
round 1 (@8d26106) claude agy grok PASS, codex NOT PASS (R1-R3); fixed in
17eba09 (docs, inventory data and the guard test only); round 2 4/4 PASS.
Full suite at 8d26106 1408 passed / 0 failed. Reports and the round-2
identity sidecars go to .agents/reports/<session>/.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Merge-workspace review of PR #46 (codex R1-R3 and C1, claude O1 and O4).
- R1: resume never attaches, and the server it exports does not reach the
calling shell; the native herdr binary does not route on environment
variables. The attach advice and the Verification block now name the
server: `herdr --session <server> agent attach <name>`, with <server> the
row's herdr_session.
- R2: --dry-run in resume and create skips the spawn and the registry
update, but the checks before it still run and on macOS remove the
quarantine attribute from the agent binary. The option rows say so.
- R3: test_no_deleted_body_came_back compares every current block with the
deleted bodies by exact sha and by a whitespace-insensitive fingerprint
(stored as base_fingerprint), independently of the inventory, so a
deleted block cannot return by being listed again under class c.
Replaying the review's probe (the deleted resume Workflow re-added and
inventoried as c) fails the new test and passes the old one.
- C1 / O1: the base create call block at 263-266 (class c) folded into the
Workflow call is recorded under "superseded"; the 37 base blocks are all
accounted for.
- O4: create exit 42 also covers MAM_KILL_ON_READY_TIMEOUT=1; stop's no-op
exit 0 does not apply with --purge-conversation.
- The inventory's positions and shas are regenerated for the edited blocks;
every classification is unchanged.
No product code changed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Stage 1: tests/data/skill_code_blocks.json lists all 37 fenced blocks of the
eight SKILL.md files at base 45d1621 (indented fences count), classed (a)
duplicate of a script, (b) procedure no script offers, (c) example/call/
diagram, with script file:line evidence and recorded mismatches.
tests/test_issue42_skill_blocks.py fails when a block changes without the
inventory, an (a) block comes back, a removed block's entry call is missing,
an exit-code table drifts from the script, or evidence points nowhere.
Stage 2: the five (a) blocks are gone - resume Workflow, stop pre-flight,
create pre-flight / Workflow / manual YAML append - replaced by the
entry-point call with option and exit-code tables. Each exit table has a
catch-all row: under set -e a failing sub-step's code passes through.
create's exit 4 is real (the YAML write raises SystemExit(4) for a duplicate
running row) and is listed as such. Doc-to-script mismatches are fixed in the
doc (resume Case 3 / no attach, stop --purge-conversation --yes, status
--json keys, create cmd_full and the nonexistent yaml-edit script). No
behaviour change, no version bump.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
PR #41/#43 review follow-ups F1-F6, implemented by creator-claude-01 in a
loop and reviewed by claude, agy, grok and codex: round 1 (@95eabc6)
codex NOT PASS (a failed env load cached as loaded), fixed in 75f364e;
round 2 4/4 PASS. Full suite at 75f364e 1380 passed / 0 failed. Reviewer
reports (round 2 with identity sidecars) and the creator's reports go to
.agents/reports/<session>/.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Issue #26: one .mam.env parser for the wrapper and the Python tools,
decimal integer knobs, tests on a local MQTT broker. Merge-workspace
review: round 1 (@b58edac) claude agy PASS, codex NOT PASS (R1-R4), grok
no verdict (its late report, archived as report-c9ed5b14.md, found R1 too);
round 2 (@0eec795) 4/4 PASS. Reports and round-2 identity sidecars go to
.agents/reports/<session>/.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Issue #33 (job budget, wall_clock_timeout vs deadline_timeout) and
Issue #30 (notify-only stall watcher). Reviews of PR #41 in the merge
workspace: round 1 (@1673818) claude N1 / codex R1 R2 NOT PASS, agy grok
PASS; round 2 (@af0a2a2) grok N1 / codex D1 D2 NOT PASS, claude agy PASS;
round 3 (@9200562) 4/4 PASS. Reports and the round-3 identity sidecars go
to .agents/reports/<session>/. The design consensus on items A-D is in the
VERSIONS.md entry.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Issue #34: the loop waits for the creator job's sub-jobs, reviewers get
the sub-job table as evidence, a run with sub-jobs voids a round whose
source changed during review, and a pane caller reaches an agent of
another workspace on its own server. VERSIONS.md carries the entry; the
merge-side reports (round 1 @ed293dc and round 2 @e80186f with identity
sidecars, base ed293dc -> head e80186f, evidence review-tree) are archived
under .agents/reports/<session>/.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
- registry: parent_job field, register --parent-job fails closed on an
unknown parent, children_of/descendants_of, list --parent
- wrapper: --parent-job; $MAM_JOB_ID default (ignored with a warning when
this registry lacks it); brief line and one sentence at the end of the
injected instructions
- create_session: the onboarding instructions carry the same sentence
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Issue #36: delegated prompts travel by file, a delegation that fails ends
the loop with rc 1. VERSIONS.md carries the entry; the merge-side reports
(round 1 @59be1c8 and round 2 @708262a with identity sidecars, base
59be1c8 -> head 708262a, evidence review-tree) are archived under
.agents/reports/<session>/.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Review of f26952c (codex, job cc849cbf):
- P2: create_session registered the task by file but then pasted the whole
task into the onboarding text, which send_keys_safe hands to `herdr agent
prompt` as one argv string. A 130,500-byte --submit-job was registered
intact and never reached the agent (execve E2BIG). The task now goes into
<jobs>/<id>/brief.md, as the delegate-job wrapper does, and the injected
text names that path. The publish command and its dispatch token (#23) are
unchanged. The other send_keys_safe callers (the wrapper's short pointer,
stop_session's exit key) carry no task text.
- P3: delegate_job_safe's write-failure branch left the partial prompt file
behind and logged to stdout, which callers capture. It now removes the file
and logs to stderr.
S8 (real create_session.sh, 100 and 130,500 bytes, strace when present) and
S9 (the real function, RLIMIT_FSIZE write failure) carry the reviewer's
reproductions over; all three fail on f26952c.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Issues #23 and #24: cross-server herdr scope, agent-agnostic start-up
dialogs, onboarding delivery judged by the agent's own started event,
verified rollback. VERSIONS.md carries the entry; the merge-side reports
(round 1: claude/agy/grok; round 2 on 05b0ea4: all four, with identity
sidecars, base f4ba205 -> head 05b0ea4, evidence review-tree) are archived
under .agents/reports/<session>/. The round-1 codex report was lost to the
host sleeping mid-review; its two probes are recorded in the changelog.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Two probes from the merge-side codex review (its report was lost to an API
outage; the probes and their results were recovered from its transcript):
- create_session.sh and resume_session.sh set AUTO_TRUST=1 and then
exported MAM_AUTO_TRUST from it, so the documented MAM_AUTO_TRUST=0
without a flag reached handle_startup_dialogs as 1: the operator had
turned automatic trust off and it stayed on. Now AUTO_TRUST defaults to
the environment value and an explicit --auto-trust/--no-auto-trust wins.
- _mam_dialog_pick exited on the first dialog inside its resend cooldown
instead of looking further down the screen, so a skipped (--no-auto-trust)
or just-answered trust prompt hid the codex update notice below it, which
the docs say is always answered. It continues to the next dialog now; S5
and S5b (one send per cooldown, one retry after it) still hold.
Regressions: S6c, S5f (startup path), R3b and C3 (resume/create, env only,
flag over env). Each fails without the fix. The create skill doc names the
environment form and its precedence.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
Trial 2 of this branch: cross-server delivery was restored (f7cee00), the
agent completed its onboarding and published started and completed - and
create_session still waited 180 s, called the onboarding unconfirmed, and
tried to write `error` over the completed job (refused by the registry).
delegate_wait_started required the started line to carry data.dispatch_token
(review R1: a stale attempt's started must not vouch for this one). But
publish_event only writes that field when its command line carries
--expect-dispatch-token, and create_session's hand-rolled onboarding command
never did - the wrapper's _pub_cmd does, which is why every wrapper-dispatched
agent's started has the token and no create_session onboarding ever could.
Five reviewers missed it because the fixture wrote the token into the started
line by hand: a fake producer that differs from the real one.
Fix, along the wrapper's existing mechanism rather than a new one:
* delegate_publish_cmd (lib.sh) builds the publish command for an agent, with
--expect-dispatch-token and --attempt-id, and create_session uses it. The
agent's own started now names the dispatch, and publish_event's CAS check
refuses a stale dispatch's event at the source. S7l pins the builder
argument-for-argument against the wrapper's _pub_cmd.
* delegate_started_recorded (lib.sh) is the one definition of "this
dispatch's started is on disk" - JSON-parsed, token-matched when a token is
given - used by delegate_wait_started and by the wrapper's _wait_started in
place of its regex. The wrapper keeps its status and subscriber routes
(remote agents); the shared part is the evidence contract.
* An already-terminal job is left alone when the wait expires (req. 4): no
error over a completed/error/cancelled record.
The fixture now registers the record through registry.py and publishes
started through the real publish_event.py; the only synthetic line left (S7h)
exists to pin key-order independence and says so. S7i pins the trial-2 defect
as the negative it is (old command shape -> started without token -> not
acknowledged); S7j runs the builder's command through the real producer and
is acknowledged on the first probe; S7n shows the wrapper accepting the same
real line and rejecting a stale token; S7m pins no-write-over-terminal. S7f
moves from "publish error" to "leave the terminal job alone", per req. 4.
Option B (verify data.hmac_sig with the job's auth_token, plus record-token
still equal) was weighed and not taken: verify_hmac returns True with no
auth_token - PoC mode, which the trial and every sandbox run in - so it proves
nothing there; in production hmac_sig authenticates the signer but does not
identify the dispatch, so R1 would rest on a current-state check that a stale
line plus an unrotated token passes; and it would diverge from the wrapper
unless the wrapper changed too. Token exposure is not new: every wrapper-
dispatched instruction already carries the dispatch token in the pane, and it
is a per-dispatch CAS nonce, not the HMAC secret.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Two defects in the started-based verdict, both found by the full suite.
The first was destructive. No `started` within T exited 1, which fired the
rollback trap and killed the session. That is worse than the bug being fixed:
create_session already has a rule for "the agent is not responding but the
process is alive" one step earlier - a TUI that never signalled ready is
preserved for diagnosis - and an onboarding that is merely slow is the same
situation. The job is marked `error` and warned about loudly; the session, its
row and the pane stay. MULTI_AGENT_RULES §6 has the GM gate on this job
anyway, so an unacknowledged onboarding still blocks real work.
It also broke 21 tests across 8 modules (create/sanity/tier2/tier3/tier4/
issue15/issue19/issue21) and added roughly an hour to the suite: no sandbox
has an agent to publish `started`, so every create_session test sat out the
full 180 s and then failed on the rollback. MAM_ONBOARD_ACK_S shortens that
wait, and the sandbox fixture sets 3 s. It is deliberately not the wrapper's
MAM_INJECT_ACK_GRACE_S - that is a different wait whose own tests assert on
the value they pass, and reusing it made issue18's acceptance2 read "within
3s" instead of "within 4s".
The second: the delivery failure published `error` and then the rollback
published `cancelled` on the same job, which the registry refuses
("conflicting cancelled after terminal error"). Terminal events are published
exactly once, so the trap now skips a job this run already closed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Plan commits 3 and 4 landed together: both rewrite the same region of
create_session.sh and share one test file, and splitting them afterwards would
have meant hunk-level surgery on already-verified code.
Delivery verdict (item 3). create_session --onboard called a healthy agent a
failure three seconds after registering the job: inject_instructions ->
send_keys_safe returns rc 4 when `agent prompt --wait --until working` fails
and the pane is still not working two seconds later, which is exactly where a
TUI that is still starting up lands. It never got the ISSUE-18 confirmation the
delegate-job wrapper has. rc 4 is a warning now and the agent's own `started`
decides, within the job's own ack timeout; rc 1/2/3 stay hard failures, since
they say the pane never quiesced, a dialog was blocking, or the text never
appeared, and no later `started` repairs those.
The self-published `started` had to go first. create_session published it on
the agent's behalf right after injection, so the registry showed a started job
whether or not the agent ever saw the prompt - with that line in place, no
started-based verdict can mean anything.
delegate_wait_started is lib.sh's counterpart to the wrapper's _wait_started
(plan item D10-b). The wrapper's own `submit` was the first choice, but it
blocks until the job reaches a terminal state, and create_session must return
as soon as delivery is confirmed; the on-disk evidence it reads - registry
status past `pending`, or a started line carrying this dispatch token - is
written by publish_event before its wire publish, so no broker is involved.
Rollback integrity (item 4). The trap ran `_herdr kill-session -t
"$SESSION_NAME" 2>/dev/null || true`, so a kill that closed nothing still
printed "rolling back" while the session kept running, and the row stayed
`running` pointing at the job the same rollback had just failed (#23 symptom
5). With item 1 in place this check is finally trustworthy: the kill used to
fail because the caller's workspace id hid the pane.
The trap now reads the kill's rc, confirms with has-session, closes the
onboarding job as `cancelled` (a terminal transition the registry knows, and
the truthful one - the job was never dispatched), and marks the row. A session
still alive exits 43 with the manual cleanup command, distinct from 42 (TUI
never ready).
The row keeps a schema-valid status rather than the plan's `error`, which
atomic_yaml.py rejects: `terminated` once the session is really gone, and
`running` when the kill failed, because it is still running. What makes that
case legible is the note, the cleared delegate_job_id and exit 43 - not a
status claiming the opposite of the truth. The row is never deleted; a deleted
row takes the evidence with it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
handle_startup_dialogs was a claude-only string ladder called only when
AGENT=claude, and wait_for_tui_ready does not treat a dialog screen as ready.
So agy, grok and codex sat on their directory-trust prompt for the full
readiness timeout, the session was recorded as tui-ready-timeout, and
onboarding injection was disabled (create_session.sh:313-314) - a human had to
accept each dialog by hand.
The signatures and accept keys move to the adapters as `startup_dialogs`
(pattern, keys, kind), published through the existing `facts` eval contract as
MAM_STARTUP_DIALOGS. claude's ladder moves over verbatim; agy takes Enter, grok
y, codex Enter, and codex's update notice takes 2 (Skip), which blocks a start
just as hard while changing nothing. kind is `trust` or `notice`:
--no-auto-trust leaves trust dialogs to a human and still answers notices,
since trusting a directory makes codex load its project-local config, hooks and
exec policy.
Re-sending is gated by time (MAM_DIALOG_RESEND_S, default 5s), not by frames.
#24's comment reports that `--source visible` keeps showing an accepted dialog
for a while, and _pane_capture hardcodes that source, so "still on screen" is
not evidence the keys were missed; acceptance is judged by the ready tokens
alone.
Two existing tests asserted the old shape and are updated, not weakened:
b19's "Escape branch stays" now looks for it in claude's adapter, and c1b's
"Enter once per iteration" becomes "never twice for one frame", which the
cooldown satisfies strictly more tightly.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Issue #25: a promoted plan is reused only with --plan-file; the implicit
newest-plan load is gone in both runners, and an empty planner report no
longer passes silently. VERSIONS.md carries the entry; the four merge-side
review reports and their identity sidecars (base 6fc9194 -> head 1e82651,
evidence review-tree) are archived under .agents/reports/<session>/.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
- worktree_prepare.sh installs the same files deploy/install.sh does
(.agents/skills/multi-agent-mux-delegate-job/requirements.txt: paho-mqtt,
PyYAML) plus requirements*.txt. Without paho the loop wrapper, stop_session
and the subscriber tests fail in the worktree - seen as 4 suite failures
in a worktree without its own venv, all passing with one.
- review D1: a failed pip install removes the venv it half made, so the next
run recreates it instead of reporting VENV=kept (test w1c).
- review D3: the [worktree context] block's uncommitted-changes flag uses
--untracked-files=all like the guards.
- review D2: the create skill doc says the install needs the package index.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
R1/R2: `git status --porcelain` honours status.showUntrackedFiles, so a
`no` setting in a submodule (or the worktree) hid untracked files from the
checks that justify worktree_remove.sh's single --force and from the review
snapshot guard (_dirty_submodules), which then approved content it never
captured. All three checks (lib_py/worktree.py, worktree_remove.sh,
runner._dirty_submodules) now pass --untracked-files=all. The source tree's
own snapshot was already safe: `git add -A` ignores that setting.
R3: worktree_prepare.sh no longer links .venv to the control root's. It
creates the worktree's own venv with the control root's interpreter and
installs requirements*.txt; a leftover link is refused (exit 1), a control
root without an interpreter is a warning (VENV=skipped); VENV= is printed.
Tests: w1 (real venv, kept on re-run), w1b (link refused / skipped), w6e
parametrised over the hidden setting for the post-stop recheck, w6e2
(hidden untracked file in the submodule and in the worktree blocks the
removal), l17 untracked-hidden / during-hidden (guard refuses before the
first review; a mid-round hidden file makes the round stale). Each new
case fails without the fix.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VCgu9Fa9iT7TbJUSzdq4Ff
- N1 (planner-reviewer-claude-01) worktree_remove.sh: when `git status` fails
in the worktree (e.g. its directory was deleted), judge_tree returned before
METHOD was set and the summary line died under set -u with exit 1. METHOD
now has its default before the function; the refusal is exit 3 with
"nothing was changed", plus a `git worktree prune` hint when the directory
is gone. Same for --dry-run.
- R4 (planner-reviewer-codex-01) worktree_prepare.sh: a submodule path that is
not initialised yet but already holds files made the clone fail, and the
rollback's `deinit -f` then deleted those files. --submodules now refuses
such a path up front (new exit 3, nothing changed), which also makes the
rollback's premise true: after a failed attempt, everything under a
not-yet-initialised path was created by that attempt.
- R5 (planner-reviewer-codex-01) lib_py/worktree.py: `submodule foreach`
visits initialised checkouts only, but a deinit'd submodule keeps its
repository - stash and unpushed commits included - under the worktree's
modules/, which the forced removal deletes. Every repository under modules/
(nested ones included) is now checked directly (`--git-dir`): commits on no
remote (HEAD included) and refs/stash with its reflog. A directory that looks
like a partial repository cannot be checked and refuses.
- Also: the rollback loops are C-style (`seq 0 -1` counts down on BSD), and
prepare warns that --base is ignored for an existing branch.
Tests W6i (deleted worktree directory, with and without --dry-run), W6j and
W6j2 (stash / unpushed commit in a deinit'd submodule's repository) and W6k
(existing files in an uninitialised submodule path). Skill docs and the design
doc describe the new refusal and the wider repository check.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017R3Eg7MEpSWnm1oB9ZPP6G
Review ffb5055e (planner-reviewer-codex-01):
- R1 [P1] worktree_remove.sh: the clean-tree and submodule checks ran only
before --stop, so work a stopping session wrote (or committed, or stashed)
was deleted by the forced removal. The checks are one function now, run
before any change and again right before the removal; any finding then
refuses with exit 3 (sessions already stopped, worktree kept), and the
removal method follows that latest look.
- R2 [P1] lib_py/worktree.py: a submodule stash is reachable from no branch
and leaves the status clean, and the forced removal deletes the worktree's
own submodule repositories. `git stash list` in every submodule is a
separate finding (`stashes`), refused like dirty/unpushed; errors fail
closed as before.
- R3 [P2] worktree_prepare.sh: when one submodule cloned and a later one
failed, the rollback deleted modules/ while the cloned checkout's .git file
still pointed into it, and `git status` failed in the worktree. The rollback
now undoes only this attempt: each submodule whose git dir did not exist
before is deinitialised before its git dir is deleted; submodules
initialised earlier are kept; shared submodule.* config values are restored
exactly (new keys removed, pre-existing values put back - init and deinit
both write .git/config). A final `git status` either confirms a usable
checkout or prints what to inspect.
Tests W6e (a write during stop keeps the worktree and the file), W6f (a
submodule stash refuses), W6g (partial failure on a fresh worktree: usable,
no modules/, no new config, then prepare and a plain removal work) and W6h (a
failing retry keeps the earlier `sm` init and its config value). The skill
docs and the design doc describe the recheck, the stash rule and the scoped
rollback.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017R3Eg7MEpSWnm1oB9ZPP6G
- design doc §3: the hand-assembled worktree/.venv/create recipe is replaced by
worktree_prepare.sh --creator-agent (its BASE= line is the base to record);
cleanup is worktree_remove.sh --stop; resume needs the original path.
- §4: what the helpers do - no .mam.env copy, the shared info/exclude, opt-in
submodules (any init leaves modules/, which blocks a plain remove), removal
of submodule worktrees with one --force after the checks, and never by hand.
- §5: item 4 is loop A on this branch; phase-1 items are done in v4.6.0.
- §7: create (control-root reconcile/watchdog, loud binding failure), resume
(original cwd only, binding persisted when already running), stop (the
removal helper's checks), monitor (register via create/prepare only).
- create / stop / resume SKILL.md: usage and exit codes of worktree_prepare.sh
and worktree_remove.sh, the original-cwd rule and its exit code (1 -> 2).
Frontmatter versions are unchanged.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017R3Eg7MEpSWnm1oB9ZPP6G
- worktree_prepare.sh (create skill): idempotent `git worktree add` next to the
control root (a path inside it is refused), .venv linked to the control
root's, /.venv and .pytest_cache/ in the shared info/exclude, submodules only
with --submodules - a failed init removes its own traces (modules/, new
submodule.* keys in the shared config) and exits 1, since git refuses a plain
`worktree remove` once modules/ exists - and an optional creator registered
in the control root's .mam with MAM_NO_SUBSCRIBER_DAEMON=1. Prints BASE= for
`run_loop.sh --worktree ... --base ... --review-only`.
- worktree_remove.sh (stop skill): every check first, nothing changed on a
refusal (exit 3): linked and not locked, running sessions only with --stop,
no open job of those sessions except the delegate jobs stop cancels, clean
tree including submodules, and inside submodules no modification and no
commit on no remote (HEAD included - submodules sit on a detached HEAD).
Then stop, re-check jobs, and remove: plain, or a single --force when the
worktree holds submodule state.
- lib_py/worktree.py: the parsing behind both (stdlib only).
- A8: create's immediate reconcile and watchdog use the control root's code
and WORKSPACE_ROOT, not the worktree --workspace's copy.
Tests W1-W6d (local submodules allowed through GIT_CONFIG_* env only).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017R3Eg7MEpSWnm1oB9ZPP6G
GodopuandClaude Opus 5.5
GodopuandClaude Fable 5.1