Thanks @Agapenordev! Supervised browser hand-off now waits for process-group/Job Object shutdown, and private-profile removal retries transient locks for up to five seconds. Missing profiles count as already cleaned.
Reviewed source head 6e3398d28b. Hosted Ubuntu CI is green. On AnayPC, the applicable engine/platform tests, clippy, layers, locks, secrets and wasm checks pass. Two Windows UI-automation tests also fail on pristine main because the SSH session cannot create interactive WinForms windows; their baseline and candidate failures are preserved.
Related #810 remains open: this follow-up does not establish that the broader live macOS/Brave issue is fully resolved.
A browser or Electron app we launch gets its private profile folder removed when it closes or fails to start, after its process is stopped; a removal that fails is logged. Refs #810.
The builder's plan card can be folded away, and the fold is kept in the project's feed until the builder makes a new plan (refs #504).
Supersedes #779.
Co-authored-by: Fábio Soares <fabiodev1123@gmail.com>
A build whose checks keep finding no working build stops after 10 in a row, across Try again presses, until a new request starts the count again (kept in craft.json as not_working_checks), and with app/ empty the builder is asked to write the project and show its build's output instead of fixing a build that doesn't exist (refs #736).
Supersedes #776.
Co-authored-by: Fábio Soares <fabiodev1123@gmail.com>
Chrome, Brave, Chromium and Edge installed as Flatpaks on Linux are found, and a Flatpak or Snap browser keeps its private profile in the home folder its sandbox shares, so it can start and its port can be read (fixes#747).
Parser follow-ups to #722: Java reference searches keep earlier hits and report partial coverage, DEX signed LEB128 takes at most five bytes, legacy AAPT long strings recover within a shared budget, and ext indirect metadata is read within bounds.
A person's own project, edited in place with nothing to recreate, is checked by its own build and then its own tests, in any language a toolchain table knows (Rust, Node.js, Python, Go, .NET, Maven, Gradle, Make), instead of the checks made for a recreation (no --screenshot option, self-test or MECHANICS.md added to their code). A project with no tests, or one Craft Anything doesn't know how to build, isn't a failure and the note says so.
A project edited in place (Work on my project, or rebuild --existing edit) keeps the stack its own files use (an engine's project file, then Cargo.toml or package.json dependencies, then other project files), unless the request names another; a Godot folder no longer gets Unity from a request that reads as a game.
The pasted-picture test accepts a third paste that lands in the same millisecond as the first and takes its freed name, and checks the first paste's content is gone. Test only.
A picture over 20 MB is refused when it is added (prompt box, drag and drop, Make's picker) with a message, and copy_pictures refuses the whole list before copying anything, instead of dropping it silently when the message is sent.
GitHub Copilot CLI can be a coding agent over ACP, with its models chosen through the shared agent model catalogue; options that could bypass the app's tool restrictions are refused, and agents share investigation findings safely.
Auto mode's delete check compares paths as written and reads quoted or escaped paths with spaces as one argument, so it neither refuses deletes inside a project with capitals in its path nor lets a delete of the folder beside the project through. Replays #357.
Co-authored-by: samurai <samuraideng@gmail.com>
Claude Code chats get no built-in tools and builders an explicit list (commands, files, search, todos), so tools that message other sessions, publish or schedule work never start; in Auto mode a tool without an approval rule is refused (fixes#635).
Replays #576, #577, #581, #589, #624, #625 and #628. Builder steps stop after 25 failed builds in a row, OpenRouter tool requests ask for a provider that supports tools, a start that never opened a window draws in software next time, spinners stop redrawing every frame, local builders get a lean tool list, lldb call observation works and reports failures, and LM Studio is found on other ports.
Co-authored-by: Fábio Soares <fabiodev1123@gmail.com>
Pictures can be pasted into the prompt box with Ctrl+V (Cmd+V, Shift+Insert): a copied image is saved as a PNG in the person's own cache folder and attached, copied picture files are attached as they are, and attachments show as picture tiles.
JavaScript analysis reads a source map comment from the last eight lines of the whole script and keeps its full URL, so originals are restored from large inline source maps.
Builder file tools and the approval policy follow links the way the file system does, so a symlink or junction inside a project can't lead a builder's read or write outside it (CWE-59); build_files_list shows links without walking into them.
web_scripts_save copies only the scripts that fit its max_scripts and max_bytes; scripts past them are listed by URL only, so a large or hostile capture costs no more than the export's own limits.
Refs #44, #110 and #481. A build stops after two builder turns in a row change no app files, counted across runs in a versioned inbox/build-progress.json; a new request or a real edit starts another attempt. File digests are cached by size and modification time, so unchanged files aren't read again.
Refs #498. A tool argument of the wrong type is answered with its JSON Pointer, the type error and how to correct the call, so a model can fix it; schema types stay strict.
Fixes#520. Supersedes #521. core_archive_extract skips ZIP symlink, device, FIFO and socket members and names each in skipped, instead of writing a symlink as a file holding its target.
Co-authored-by: NetterSilver <133535268+Silvertree2010@users.noreply.github.com>
Supersedes #216. A prompt brought back with Up keeps the pictures it was sent with. feed.json moves to version 3: a User item with pictures is saved as {text, pictures}, one without stays a plain string, so older files read as before. Picture paths are absolute, so they don't come back after the project is moved.
Co-authored-by: muraa-p <muraa-p@users.noreply.github.com>
runtime_captures_compare's left and right parameters are declared as a string or an object instead of a boolean schema, which made Claude Code drop every tool from the MCP server; on the command line a JSON object passed to them stays an object.
Fixes#540. Codex's codex exec can't ask for approval, so setup and the builder approve rebuild_start (which asks the person itself) on Codex's side, keeping the person's other settings in the entry; a refused call explains the fix instead of showing null.
Fixes#518. With if_tampered=continue, core_archive_extract writes a ZIP member whose CRC does not match (the mismatch stays in contradictions), as its remediation says.
Fixes#519. core_archive_show counts nested archives that a depth or size limit left unopened, reports a nested_not_opened limitation and marks coverage partial.
Fixes#516 and #517. Archive extraction refuses member names with a colon or a trailing dot or space, which Windows would change (merging two members, or hiding bytes in an alternate data stream).
Fixes#510. The built-in engine lists DNS answers like tshark does: each A/AAAA address keyed by the response's first question, so CNAME chains end at the name asked. Both engines report the dns_answer_attribution limitation.
In Deep mode, Ghidra's pseudo-C uses the names set with native_name_set in the function field, the summary and the code (whole identifiers only, never inside comments, literals or fields); names that can't stand in C or would clash are listed in a limitation.
A stand-in backend no longer replays a recorded 304 without its cached body: such samples are kept for diagnosis, skipped in replay, and a request with no usable answer gets cached_body_unavailable. Builds count them.
Fixes#335. Before a tab is clicked, web_page_states_capture marks the tab its list has selected and clicks it again after the picture, so later states and the next capture see the page as it was.
Fixes#469. craft.json records which coding agent saved the session (agent_session_by). A build or a question continues the session only with that agent; after the builder changes, the new one starts from the brief. A follow-up to an agent with no saved session also gets the brief.
settings_set keeps a test session's engines.only_configured across its settings round trip, so a test that changes a setting still doesn't find this computer's engines. New testing::live_session() finds installed engines for the --ignored live tests; the live website tests use it.
Fixes#339. web_page_states_capture drops the data-craft-state marks an earlier call left before marking, so a capture at a second width opens that width's own controls instead of a hidden one from the first width.
Fixes#451, fixes#452, closes#446. Claude Code's result for a --resume session it doesn't have is recognised (StreamParser::session_missing); the build forgets the stale id and starts a new session from the brief, what is in app/ and what the project kept of the old conversation, and questions answer from the project's files. After every builder turn the project keeps a private copy of the session, which is put back under Claude Code's folder on another computer when it isn't there (never over an existing one).
Fixes#336. web_site_capture keeps the first 4 MiB of the page's HTML; when the page has more, the evidence carries an html_cut limitation with both sizes and the README says page.html is the first part, cut at the size limit.
Test sessions set engines.only_configured (never saved), so Settings::ghidra_launcher and engine_path look only where the settings and the managed store say, not on PATH or in the usual install folders. Tests now pass the same on a machine with Ghidra installed. New locate_configured and ghidra::get::headless_configured are the first half of locate and headless, which return the same as before.
native_function_find works out the offset as a signed value, so an address in a block below the function's entry gives a negative offset ("-0x70") instead of wrapping to a huge number or panicking with overflow checks.
Merges the whole SQLite stack in one commit (#410, #411 and #412): WITHOUT ROWID tables are read in primary-key order, rows older than ALTER TABLE ADD COLUMN show the added columns, and records over the payload cap say which values were cut.
Fixes#406, fixes#407, fixes#408.
Fixes#294. walk_dir returns an error when the root folder can't be listed and keeps each unlistable folder below it; core_archive_show and js_app_map mark those folders, add an unlisted_folders limitation and report partial coverage.
core_archive_show records max_entries and max_total_bytes, and a shared recorded_inventory helper checks a folder's freshness at the scope it was recorded with, so core_archives_compare and the Android, Apple and Harmony graph tools no longer refuse unchanged folders as file_changed.
game_unity_bundle_list adds a files_unread limitation when a bundle has more serialized files than it reads (16) and says "at least N object(s)" in its summary.
In summary mode formats_ai_weights_show saves every tensor in its evidence record with complete coverage, so core_evidence_browse can page through all of them instead of stopping after the first page.
Declared source maps on the script's own loopback origin (http://127.0.0.1 or [::1] with a port) are now resolved against the script URL and fetched with the bounded HTTP client, reporting fetched, bytes or the reason. Other hosts, ports, schemes and remote https maps are still never requested.
app_folder_explain counts files it couldn't open to look inside (unopened_files) and reports partial coverage then, and when its walk couldn't read something, instead of a complete 'nothing points to a runtime'. Fixes#292.
Frozen and light pins keep a project's runtime models in assets/obj; only obj folders beside a .NET project file are left out as build output. Fixes#274.
Agapenor