* fix(hopper): keep installed apt dependencies at their versions during Linux setup
apt-get install upgrades every listed package that is already installed, so Linux setup could upgrade the system python3 and X11 libraries while installing Hopper. Runtime dependencies are now installed with --no-upgrade in a separate transaction before the Hopper archive, which may still upgrade. The setup approval plan lists both commands.
* fix(hopper): narrow the apt upgrade claim and cover split transaction failures
* fix(evidence): preserve JSON ownership through metadata updates
---------
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
* fix(setup): materialize the pinned package-runner cache before the first client start
Package runners key their package cache by the requested spec, so the
entry installed for the @latest invocation that ran setup cannot serve
the pinned rea-agents@<version> registration. The first client start
installed that entry inside the client's MCP startup deadline and timed
out (three consecutive 30 s timeouts observed in one client log; a
controlled cold install with every tarball already cached took 22 s).
Setup now materializes the pinned entry once after applying
registrations: best-effort, bounded by a timeout, launched from a
neutral working directory, and non-fatal on failure. The setup result
carries the prewarm summary, interactive output prints a warning when
the step fails, and direct launcher registrations keep skipping it.
Fixes#7027
* fix(setup): gate prewarm to applied registrations and supervise the runner tree
Address review findings on the prewarm step:
- Gate the prewarm on at least one applied client registration
(configured_<client>), so an empty plan or a skill-only run never
installs into the package-runner cache; the step now only follows
registrations the user approved in this transaction.
- Replace execFileOutput with a supervised spawn: on timeout the whole
runner tree stops through cleanupWindowsProcessTree on Windows and a
process-group kill on POSIX, because killing only the direct child
leaves npm holding the inherited output pipes past the deadline.
- Emit started and completed progress events around the prewarm so an
interactive cold install no longer looks like a frozen setup.
* fix(setup): forward terminal interrupts to the detached prewarm runner
A detached POSIX package-runner group no longer receives the terminal's
Ctrl-C, so setup could exit while the runner kept installing in the
background. The supervised launch now registers SIGINT and SIGTERM
forwarding for the duration of the prewarm: the runner tree is stopped
through the same family cleanup as the timeout, and the default
termination is restored afterwards. An interrupted launch reports a
failure reason instead of a hang.
* fix(targets): keep original DOS padding out of e_lfanew
A complete DOS header with a nonempty relocation table that ended before
e_lfanew treated the padding dword at 0x3c as a Windows declaration. Without
a recognized Windows signature, fall back to the validated DOS structure so
original DOS images are not refused as malformed Windows targets. Preserve
unresolved-candidate handling when no relocation table delimits the padding.
* test(ghidra): cover DOS padding and unresolved header candidates
* fix(targets): read Windows signatures before dismissing DOS padding
* docs(targets): reconcile DOS padding admission contract
Document that relocation-table padding inside a complete DOS header is DOS data, while an unread Windows signature beyond the current probe is requested before that rule applies.
Aligns docs/installation.md with docs/ghidra-dos.md after the MZ admission fix.
---------
Co-authored-by: dhruvac29 <dhruvac29@users.noreply.github.com>
* fix(setup): report host-owned cleanup resources alongside the primary's
SetupHost.close documents that it reports the resources a setup host still
owns, but its contract returned only a string. runSetup therefore built the
ProviderCleanupError from the primary failure's cleanupResources alone, so a
resource the host failed to release reached the projection only inside a
free-text reason while the error asks the caller to review the reported local
resources.
Return a SetupCloseReport instead. closeSetupResources absorbs the structured
resources of any cause that reports them, through
ArtifactReaderFailure.cleanupObservation, and leaves opaque failures
message-only so no resource name is invented. runSetup then unions the host's
resources with the primary's before building the typed cleanup error.
Two boundary tests read the close result directly: the skill-cleanup case now
asserts the retained backup path in the report reason rather than in a bare
string, and the shared ownership helper takes an opaque close result because it
only ever awaits undefined from either host it is given.
Validation: RED then green in src/application/Setup.recovery.test.ts, where
the new case previously received only the primary resource; 116 tests across
the seven Setup and cleanup-ownership suites pass; tsc --noEmit over the
changed set, its setup callers, and both boundary tests exits 0; oxlint
reports 0 errors; oxfmt --check is clean; verify:module-boundaries passes.
The two failures in src/setupEnvironment.test.ts reproduce identically on
pristine bded9c628 with this change stashed.
* fix(setup): preserve cleanup ownership through host reports
* fix(setup): bind finalization reports to cleanup owners
* fix(setup): retain live resources in returned cleanup views
* fix(setup): preserve completed cleanup and exact installer resources
* fix(setup): retain delegated cleanup across authority changes
* fix(setup): compose callback and source cleanup authorities
* fix(setup): retain producer cleanup failure reasons
* fix(cli): render structured setup cleanup resources
* fix(setup): separate failure evidence from cleanup ownership
* fix(setup): normalize close failures and unify lease binding
* fix(setup): preserve direct lease resource attribution
---------
Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>
* feat(ghidra): bound one procedure's decompile without closing the session
Fixes#5340
* fix(ghidra): report decompile limit as a request timeout
Keeps BinarySession from marking every Ghidra operation unavailable after one procedure's decompile reaches the limit, and names the MCP tool-call deadline in the remediation.
* fix(ghidra): budget the decompile limit and report cancellation-closed sessions
Adds REA_GHIDRA_DECOMPILE_TIMEOUT_SECONDS to the derived Codex/Kimi tool deadline and forwards it through setup. Marks a remote failure that settles after caller cancellation so the timeout projection no longer promises session reuse.
* fix(ghidra): align deadline bounds and cancellation outcomes
* refactor(ghidra): shorten deadline recovery guidance
* fix(mcp): clarify the scope of inferred decompile deadlines
* test(ghidra): verify deadline workflows and document measured costs
Separate failure and registration cases by their distinct contracts.
Record actual Linux Ghidra, CLI and MCP evidence with measured context,
elapsed time, CPU and memory, and explicit remaining coverage limits.
* docs(ghidra): pin integration and distinguish baseline format failures
* fix(ghidra): own failure lifecycle wording in the provider
* test(ghidra): budget timeout fixture generation before allocation
* fix(ghidra): align capability and traversal deadline guidance
* style(config): align deadline guidance formatting
* docs(ghidra): record capability and traversal verification
---------
Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>
* fix(doctor): check only the current platform's Hopper launcher candidates
Doctor built one launcher list for every platform, so on macOS it could
report a Linux-only path as a healthy Hopper ahead of a renamed bundle,
and setup then registered that path. Give per-platform automatic
candidates one owner that runtime configuration and doctor both read,
scan renamed macOS bundles and probe the Homebrew cask only on macOS,
and share runtime's renamed-bundle scan with the doctor host.
Fixes#6228
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(doctor): keep the renamed Hopper bundle scan asynchronous
Sharing runtime's synchronous scan made the doctor host read
~/Applications with readdirSync, which blocks the event loop on a slow
or network-mounted home. Give MacHopperPaths an asynchronous scan for
doctor beside runtime's lazy synchronous one, sharing the directory
filter and candidate builder, and check that both return the same
candidates.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* test(doctor): model renamed launchers with runtime's POSIX joins
Since #5235, runtime builds macOS launcher paths with posix.join, so on
Windows the discovery table's native-join renamed launcher no longer
matched runtime's candidate. Model launchers with posix.join and keep
filesystem setup native, as MacHopperPaths.test.ts does.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Require Node.js ^22.23.4 || ^24.22.1 || >=26.12.1 across package engines, setup/doctor and the standalone installer. Validate patch boundaries and maintain runtime-independent listener identity, heap-option and HTTP schema-binding tests.
Closes#6283
Refs #6480
BREAKING CHANGE: Node versions below 22.23.4, 24.22.1 and 26.12.1 are no longer supported. Upgrade Node before installing or updating REA; the installer does not upgrade Node.
Co-authored-by: gensokyo-dev <342580529+gensokyo-dev@users.noreply.github.com>
* fix(setup): keep client and skill setup when the Hopper install fails
A failed Hopper package install returned from runSetup before any client
registration or skill install ran, so one optional provider's package
conflict (zlib-ng-compat on Arch and CachyOS) cost every other approved
change. Record the failure instead, apply the remaining plan without a
Hopper launcher, and report needs_human with the install's code and
remediation.
Fixes#5113
* fix(setup): keep the Hopper failure when a client write or the skill install fails too
* fix(setup): keep client and skill setup when the Hopper install fails
A failed Hopper package install returned from runSetup before any client
registration or skill install ran, so one optional provider's package
conflict (zlib-ng-compat on Arch and CachyOS) cost every other approved
change. Record the failure instead, apply the remaining plan without a
Hopper launcher, and report needs_human with the install's code and
remediation.
Fixes#5113
* fix(setup): keep the Hopper failure when a client write or the skill install fails too
---------
Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
* fix(hopper): keep automated Linux setup off non-x86_64 hosts
Every pinned Linux Hopper package (deb, rpm, Arch) is x86_64-only, but
hopperSetupSupported was derived from os-release alone. Setup therefore
planned and confirmed an amd64 package install on arm64 Ubuntu 24.04 and
the other tested distributions.
Classify installer support with the host architecture where the
distribution is parsed, so setup planning, preflight, installation and
package verification share one decision. Setup now takes its Linux
distribution from the doctor host, as it already does for the macOS
version, and the supported-host remediation text has a single owner.
Fixes#3419
* fix(hopper): install Linux Hopper from the setup host's distribution facts
systemLinuxHopperInstallHost read os-release with process.arch on its own,
while setup planning used the doctor host's injected architecture. Both
agree in production, but a doctor host composed with another architecture
could plan one answer and install with another.
The install host now takes its distribution reader from the caller, and
systemSetupHost passes the doctor host's, so planning, preflight and
installation share one host-fact owner. The unused default host on
installLinuxHopper, the other ambient reader, is removed.
* fix(verify-package): require Hopper readiness only where Linux setup installs it
packageSetupSupport required doctor Hopper readiness on every Linux host
whose host check passed, including installer-untested distributions and,
after the architecture gate, arm64 hosts where setup never provides
Hopper. On Linux, readiness now also requires automated installer
applicability; macOS is unchanged.
This narrows the readiness assertion from #2973 on hosts outside the
installer matrix, such as Kali and ubuntu-24.04-arm.
---------
Co-authored-by: DavidFB-creator <220631175+DavidFB-creator@users.noreply.github.com>
Co-authored-by: Kaoru0822 <208281322+kaoru0822-kitauji@users.noreply.github.com>
Fixes#6212. Parse effective argfile syntax once, admit bounded regular files with retained cleanup/cancellation ownership, freeze validated bytes for subsequent helpers and analysis launch, and share existing private-runtime writing. Resolves the continuation, FIFO and mutable-source review findings.
Verified 97 related tests (3 explicit skips), full static/docs checks, two independent local Codex reviews and 12 real installed-JDK17 grammar/property controls. Real Ghidra, JDK25 and native Windows qualification remain unavailable. Used the explicitly authorized maintainer bypass: current GitHub CI remained queued and Greptile rerun was still in progress; author-readiness metadata was not fabricated.
Assisted-by: Kogasa (gensokyo-skills:kogasa-surprise-testing)
Name REA_SWIFTC_COMPILE_TIMEOUT_MS, its effective milliseconds, the
1..2147483647 range, and the omitted 120000 ms default when a Darwin
ownership-helper compile times out. A change applies after REA restarts,
on the next ownership-helper compile.
Producer SIGXCPU and SIGXFSZ recovery says REA retains the tighter
inherited limit. Agent-visible fixed caps in this theme say this release
has no runtime override. Numbers are unchanged.
Refs #5910
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: morluto <morluto@users.noreply.github.com>
* fix(kuna): reject dangling cache roots in doctor
Treat a cache path as creatable only when no directory entry occupies it.
A dangling symbolic link, including one hidden by a trailing separator,
fails the cache check and clears provider availability while the installation
check and caller-selected registration settings remain.
Fixes#5554Fixes#5105
Co-authored-by: morluto <morluto@users.noreply.github.com>
* fix(doctor): admit the same settings native analysis rejects
Project parseConfig failures into scoped engine readiness. Invalid provider
preferences, startup timeouts, and other shared settings now fail the
selected provider and the full audit, while a JEB-only timeout stays outside
that admission.
Fixes#5454
Co-authored-by: morluto <morluto@users.noreply.github.com>
* fix(doctor): report an invalid log level from doctor
The dispatcher rejected REA_LOG_LEVEL before every command, so doctor never
emitted its readiness check. Doctor now runs and names that setting; other
commands still stop before execution.
Co-authored-by: morluto <morluto@users.noreply.github.com>
* fix(doctor): require log-level failures in every scope
An invalid REA_LOG_LEVEL stops the executable and MCP server before a client
or skill can run, so those scopes now fail readiness too. Other native
admission failures stay limited to a selected provider.
Co-authored-by: morluto <morluto@users.noreply.github.com>
* fix(kuna): strip only the platform cache separator
On Darwin and Linux a trailing backslash is a filename character. Doctor now
removes only the selected platform separator before inspecting the cache path.
Co-authored-by: morluto <morluto@users.noreply.github.com>
* fix(doctor): fail every provider on an invalid Hopper timeout
REA_HOPPER_STARTUP_TIMEOUT_MS is rejected before any native engine starts.
A selected non-Hopper provider now includes the existing timeout check.
Co-authored-by: morluto <morluto@users.noreply.github.com>
* style(doctor): format the Hopper timeout regression
Co-authored-by: morluto <morluto@users.noreply.github.com>
* fix(doctor): fail client scopes on MCP startup errors
A registered client starts the MCP server, and parseConfig rejects every
native admission failure before that server starts. Client readiness now
includes those failures, not only REA_LOG_LEVEL.
Co-authored-by: morluto <morluto@users.noreply.github.com>
---------
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: morluto <morluto@users.noreply.github.com>
Explain process-scoped startup settings in shared CLI/MCP timeout guidance and the bundled skill. Preserve Evidence and cleanup before restarting REA, reconnect MCP, and allow enough client request time.
Validated with focused tests, static/docs/package checks, real Ghidra CLI/MCP recovery, and a completed 591-second EXE analysis under CPU and memory limits.
Preserve configuration paths and constituent actions independently of aggregate status. Derive summaries from all leaf outcomes, including after final cleanup reporting, so a later failure cannot hide an earlier removal.
Closes#5079
Read shadowed registrations from the effective JSONC syntax tree before
any client is changed. Share the same typed ownership decision with the
reread before removal, including foreign OpenCode V1 fallbacks. Remove
all admitted duplicates using one member/separator planner and grant
serializer authority only for exact admitted paths.
Setup may remove its explicitly superseded OpenCode V1 path. Preserve
inactive repeated parents, unrelated entries, comments and source backups.
Use the pinned JSONC scanner instead of handwritten string decoding.
Validation: final 86 focused cases and 9 setup acceptance cases passed;
104 additional filesystem consumer cases passed. Static/architecture,
docs and discovery checks passed. A broader 186-case run had one OMP
timeout; its isolated 9-case rerun passed, with timeout cause unverified.
Controlled compiled CLI comparisons verify removal/refusal and backups;
real developer-client execution remains unverified.
Adapted from #4977 at 4cd22d4ec2.
Closes#4976
Co-authored-by: Nikhil Bhagat <knewboy.nykhil@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Verify the distribution Python embedding library during approved Linux Hopper
setup, conditionally repair the matching Ubuntu dependency, and confirm that
the library loads. Share fresh bounded observations with doctor, preserve
unknown configuration and caller cancellation, and join probe exit through
the existing process supervisor.
Validate clean Ubuntu installation, normal-user real Hopper CLI/MCP workflows,
independent ELF/SQLite oracles, cancellation and cleanup regressions, 154
focused checks and all 40 current-head CI checks. Preserve measured artifact
identity, failed attempts and local host-memory guard limitations.
Closes#5015.
Co-authored-by: Kaoru0822 <208281322+kaoru0822-kitauji@users.noreply.github.com>
Replace configuration-specific write-file-atomic delegation with a verified
input snapshot, exclusive private staging, and an explicit canonical-parent
destination. Existing files publish by rename; first-run files use an exclusive
link so a newly created destination is retained. Both consumers read back the
publication owner's selected path.
Retain descriptor and staging cleanup capabilities through recovery and host
retry. Preserve mode and ownership, remove temporary permission handoffs, and
use positional restoration writes. Keep UTF-8 diagnostics with failed cleanup.
Validation: 284 focused tests passed, one skipped across 14 files; check:fast,
format:check, docs:check (including five consistency tests), two source reviews,
and 12 completed isolated public CLI setup/uninstall cost controls passed.
Native Windows/Linux, real editor concurrency, and providers were not verified.
Final-check/publication races and same-inode concurrent writers remain unexcluded.
Fixes#4904
Replace pathname writes with the shared configuration publication owner.
Reject replacement readback and restore through the exact staged descriptor.
Retain initial-read, readback and publication cleanup owners for retry. Share
nested cleanup handling with setup without granting authority to opaque
cancellation reasons. Separate host admission from nested file cleanup gates.
Preserve uninstall file modes, including readable read-only configurations.
Hold the exact staging permission owner through writable-descriptor handoff
and failed mode restoration. Remove the superseded writeText effect contract.
Validation: 206 focused tests across 9 files, check:fast, formatting and
docs:check passed. Three uninstall replacement controls fail on baseline
7e4eed9dd while thirteen setup controls pass. Twelve completed CLI uninstall
tasks preserve equivalent outputs, config, backups and modes0400/0640.
Two read-only source reviews completed. Native Windows/Linux and real editor
races remain unverified; in-place writer exclusion is still not provided.
Closes#4902
Retain the staged file descriptor through atomic publication, readback and
recovery. Refuse detected replacements or edits and restore through that
owned descriptor instead of reopening a pathname. First-run readback failure
now retains the file for manual recovery because pathname deletion cannot
prove exact ownership. Keep both descriptor owners when nested closes fail;
opaque close rejections do not transfer ownership.
Validation: 192 focused checks in 9 files, check:fast, formatting and docs:check
passed. Three replacement regressions fail on the baseline while six existing
recovery controls pass. Six completed setup tasks preserve equivalent config
and backups. Two read-only source reviews completed on macOS arm64 / Node24.
Native Windows/Linux and real concurrent editor behavior remain unverified.
State checks do not exclude an in-place writer after the last check.
Closes#4900
Allow independently bounded Darwin helper compilation to finish before the five-second Java deadline starts. Keep Linux process-table preparation under command supervision.
Share preparation failure translation between both stages so physical compiler outcomes, exact cancellation reasons and cleanup capabilities retain one owner. Adapt contributor PR #4795 and extend preparation handoff and retry controls.
Fixes#4793
Co-authored-by: J0UH <217747670+J0UH@users.noreply.github.com>
Validate XML bytes before shared syntax parsing so literal U+FFFD remains valid while malformed bytes and declaration mismatches fail at their producing boundary.
Validation: 92 focused checks, 3 selected fake Apktool boundary checks, 6 final encoding controls, check:fast and docs:check. Ghidra coverage uses tiny real files and descriptors, without a real engine or Windows qualification.
Refs #4712
Carry producer cancellation separately from physical launch errors and diagnostic causes. Retain exact opaque reasons, LaunchSupport warnings, and rollback owners; consume the authored outcome without reclassifying failed launches from a later deadline.
Validation: focused client/provider/rollback and helper/MCP tests, check:fast, docs:check. Controlled native process ownership and Node bridge only; real Ghidra/JVM and Windows/Linux fault behavior unverified.
Refs #4517
Read bounded stable prefixes through the installation file owner, preserve unknown header diagnostics, and propagate cancellation or incomplete cleanup. Share extent and identity validation with full artifact reads.
Validated with 269 focused tests, check:fast, and docs:check. Controlled Darwin file faults reproduce the original lost owner; real Ghidra and native Windows/Linux verification remain unperformed.
Refs #4676
Reject special files without blocking, cap the properties payload at 1 MiB before allocation, and verify descriptor extent and the selected name. Preserve Latin-1 data, selected symlinks, physical filesystem facts, completed observations and exact failed-close ownership.
Reuse the stable bounded reader for profile artifacts and consolidate Ghidra file failure projection. Distinguish missing-at-open from disappearance or read failure, and attribute typed prerequisite failures at the consuming operation.
Validation: 384 focused tests across 14 files, 42 seed/extension controls, 32 final metadata/profile/CLI controls, check:fast and docs:check. Controlled macOS file/process fixtures; no real Ghidra/JVM or Windows/Linux lifecycle verification.
Bounded FIFO reproduction: baseline direct and public doctor required external cleanup around 2.2s; candidate returned independently around 0.4/0.57s, all fixture groups verified empty. Warm normal doctor baseline/candidate: elapsed 1.211/1.230s, CPU 1.239/1.247s, sampled family RSS 524/524MB, reply tokens 3116/3116. No performance improvement claimed; true cold and whole-agent costs unmeasured.
Refs #4672
Add cases where the backup copy reports success but leaves the recovery
file missing or holding other bytes; uninstall must fail before writing.
Narrow the installation guide so it promises a backup path only for the
failures that report one.
(cherry picked from commit 70bff72403)
Adolanium