Remove system strategy (legacy maps to developer). synthetic_tool still
forges the call but uses real stack text as the tool result, and now
yields ToolCall/Result events so the CLI flushes and does not glue nag
text to the next assistant stream.
Split tags: [TODO OPEN WORK] for pending/in_progress (undone work) vs
[TODO HYGIENE] for all-terminal leftover groups. End-of-turn review lists
open titles; tool docs stress open items as unfinished work.
Denylist hits ask the user (default deny, timeout) instead of only hard
rejecting when a policy hook is installed. Independent of YOLO soft-confirm;
session grants skip re-prompt for the same basename. CLI installs the hook
via install_cli_limit_hooks.
Prompt the human (optional no-echo secret) and write only to the PTY.
Tool results never include the answer so secrets stay off external APIs.
Empty/Ctrl-C cancel without writing; submit defaults to trailing newline.
Prefer $VISUAL then $EDITOR (blocking). Without either, fall back to
xdg-open / open / startfile (non-blocking) for config paths only.
/edit stays blocking-only so the temp buffer can be re-read.
Default still replaces first match only. Report how many matches were
found, how many replaced, and remaining hits so agents can raise
max_replaces or narrow old_string.
Unset/empty url still maps to durable user-data chat.db.
Explicit :memory: is not rewritten; CLI warns when not quiet.
DatabaseConfig.url default is None; engine treats None as memory.
Inject turn-start [TODO REMINDER] for non-empty stacks; rewrite end-of-turn
[TODO OPEN] copy so open items read as unfinished work. needs_review now
always true while open items remain (even if todo_* ran this turn); terminal-only
stacks still need review if untouched. Update tool docs and tests.
Create scratch dirs under system temp (tempfile), allow absolute paths
under them without rebinding the project workspace or session. Cleanup
owned temps on chat exit. Cap 16 concurrent temps; sanitize prefix.
classify_danger prompts on interactive shell/runtime launches and any -c
code (with argv + snippet preview). Confirm deny can attach a user comment
for the model. Restores ToolRegistry.execute after WIP breakage.
Add prek.toml so git commit runs ruff-pre-commit (fix then format) and
pdm-based basedpyright. Document install/install-hook; clear type warnings
that made basedpyright exit non-zero under recommended mode.
TOP = last element only for pop; multi-push puts first title on top for DFS
breakdown. Dropped nested-frame model; migrate old frames JSON to flat stack.
push of one or more titles opens a new frame (siblings at one depth);
pop leaves the frame. Supports T1,T2 → T1.1,T1.2 → pop → T2.1 pattern.
Legacy flat session JSON migrates to a single root frame.
Sub-task stack stored on session.todo_stack; model tools todo_*; slash
/todos for human ops; agent loop injects a review user message when open
todos exist and no todo_* tool ran this turn. Stack carries across /compact.
After each completed assistant+tool batch, persist and keep that prefix on
failure. Rollback only unfinished assistant/stream suffix. pending_retry
and /retry also apply when history ends with tool results (continue model
without re-executing tools).
Store provider_name/model on sessions (SQLite migrate); restore on
resume and persist on create//model//provider. Startup prefers session
selection over re-asking. ask/choose and provider/model pick use
readline with Tab completion when available.
Register REPL slash commands as a Click Group (params, UsageError, help).
Dispatch via awaitlet.async_def so sync command bodies can awaitlet()
async memory/confirm work. Completer and /help derive from the registry.
Decode and stream reasoning_content as ReasoningDeltaEvent; keep it on
assistant messages. CLI shows dim reasoning, toggles stream/verbose, and
prints full tool results when verbose.
Add process-wide prompting backends for interactive human I/O; wire CLI
limit/confirm/workspace prompts through it. Expose ask_user, choose_user,
and form_user under tools/chat and include them in DEFAULT_TOOLS.
Drop stored pending field; incomplete turns are just history ending in
UserChatMessage. Compact seeds use an assistant summary so they are not
mistaken for retryable orphans.
Keep incomplete user messages in memory/DB; roll back only
assistant/tool on failure. /retry and resume of orphan-user sessions
re-run without duplicating the user message.
After the first model call, scale char-based token estimates by
API/resolved prompt size so budget checks track near-real tokens.
/compact uses the same calibration when available.
Tool-loop rounds re-send history, so summed prompt tokens are billing
not context size. Track last_request_usage and label turn/session as
billed; clarify /status and end-of-turn lines.
NCBM