4d7c6efe9a
* feat(desktop): capture, draft, export and witness a task inside a build computer - start_build_capture runs REA's capture inside the computer through a small controller over computer-exec, after a user-space apt-get update; mark_build_checkpoint returns once the capture confirms the checkpoint; stop_build_capture returns the draft between seed and end. - export_build_environment exports with the requester's task inside the computer and fetches the bundle through data/get in bounded batches. - check_build_task grades a computer's data/get with REA's evaluator and, through seed, after_edit, near_miss and after_reset stages, records a witness with the action log and the screen recording; act_on_build_computer gains reset. - Capture progress lines carry window, checkpoint and refusal events; apps start in the home; a pack script builds the runtime archive; long tools send progress notifications. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): make build capture and witness steps retryable, and close the near-miss reset hole - A capture whose start answer is lost is stopped at once and kept unreleased until the stop is confirmed; a lost checkpoint answer is recovered from the capture's own log, and draft and export use the checkpoints its Evidence records; stop can be repeated to resume reading its result. - A witness can check its seed again after a failed read, and keeps its stages in a saving state until the record is written, so a failed save can be retried without reading the app again. - after_edit and near_miss count only click, drag, scroll, type and key, so a reset no longer stands in for the near-miss edit. - A .deb build keeps its package file for capture and passes its URL to export; bundle downloads write each checked file before fetching the next; an apt-get update failure reports its exit code and log tail; reset sends progress while it waits. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): keep the Trash out of seeds, bundle folders unwritten and first frames drawn From the blind Inkscape rebuild: - Paths in ~/.local/share/Trash get a desktop_trash reason and the shared_desktop role, so a file trashed before seed is neither owned nor seeded; the export names it in a limitation. - The keep helper runs with PYTHONDONTWRITEBYTECODE=1, so starting a computer from a bundle folder no longer writes __pycache__ into it. - start_build_computer waits, bounded, until the first frame is not one colour before returning it. From review: a .deb that apt also offers exports from apt, since export passes --package-url only when the capture says the package must come from its file. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): bound the first-frame wait's retries to the time left, with progress throughout Each retry of the drawn-first-frame wait runs under an abort that fires at the 10 s deadline, and the latest frame is returned when time runs out. connect takes the first frame itself while its progress heartbeat still runs, reporting "waiting for the desktop to draw". The wait moves to DrawnScreenshot.ts. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): serialize each build computer's calls, keep one recording id, and separate unreadable checks from failures - One queue per computer: actions, capture steps and witness checks run one at a time, so two overlapping after_edit checks can no longer both be filed and let the witness hold without a near miss. Close cancels the running and queued calls; different computers still run concurrently. - The witness keeps one recording request id per computer, so a seed retried after a lost reply recovers the recording it started. - A check that could not read its file is undecided, not a failure, so a temporary read error can't satisfy seed, near_miss or after_reset; such a stage is refused and can be checked again. - The export's file listing is written to a file and read back in bounded batches, so a large bundle no longer overruns exec's 1 MiB output. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): treat failures on cut views as undecided, require a decided seed failure at export, and keep a close's cancellation - A presence check that misses its line, row or pointed value in a view cut short is undecided rather than failed, since the data may sit in the part left out; a witness stage can no longer be proved by such a failure. - Export counts only decided failures on the seed, and refuses a task whose every seed failure is undecided. - A step whose turn was aborted by a close or the caller returns cancelled even when its work finished, so start_build_capture no longer reports success for a computer being removed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): keep a step's saved result when a cancellation arrives during its commit Reporting cancelled for any step whose turn was aborted after its work replaced a completed export or witness save with a bare cancellation, losing what was saved and leaving the witness unreadable. Steps now return what they completed; only start_build_capture, which commits nothing, checks its turn after the post-launch screenshot and reports cancelled. Asking again for a saved witness's final check returns its record. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): refuse a saved witness's repeated check for another folder Asking again for the check that ended a witness returned its saved record whatever directory came with it, so a caller naming another folder got success though nothing was saved there. A different folder is now refused, as an unfinished seed's retry already is. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(desktop): assert the facts of cut-view and undecided outcomes, not their sentences The new cut-view, missing-value, undecided-refusal and witness tests froze whole descriptions, so a harmless rewording broke them. They now check each outcome's id, passed and undecided exactly and the facts its message must carry (file, pointer, line, row, stage) separately. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(desktop): read XML with a plain DOCTYPE, and check an element's whole text - Both readers accept a DOCTYPE with only a name and an optional PUBLIC or SYSTEM identifier, record it as the view's doctype, and never fetch the DTD or expand entities beyond XML's five; an internal subset, an external entity and a parameter entity stay refused alike. - text_equals and text_contains compare the concatenated text of the element (or text node or attribute) a pointer reaches, as XPath string() does, with optional normalize_whitespace; both evaluators agree, and a text check on a truncated element is undecided. - Whitespace-only text between elements is dropped only when the element holds no other text, so text split into runs keeps its spaces. - The draft suggests text_equals for an element whose text changed. So an AbiWord .abw reads as XML, and its paragraph's text is checked whole however AbiWord splits it into <c> runs, as Writer's spans are. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): draft whole-text checks on the paragraph, read only text as text, and skip asctime stamps - A changed value is read at both ends, so text that became a run still marks its paragraph, and the drafted text_equals moves up out of mixed content to the nearest ancestor named by an attribute, such as AbiWord's p by its xid. - Text checks accept only a text node or an attribute value; element names and doctype fields fail, in both evaluators. - isTimestamp also recognises asctime dates (Www Mmm dd hh:mm:ss yyyy). - XML files are drafted with the keyed diff the archive entries use, so no suggestion reaches an element by position; the shared logic moves to appTaskDraftNamed.ts. - Text-check tests assert facts, not whole sentences. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): keep checks on a changed attribute inside a paragraph whose text check replaces its runs' A drafted whole-text check on a paragraph hid every change inside it, so a link whose label and href both changed drafted only the paragraph's text and the new destination went unchecked. The text check now covers text and runs added or removed whole; a changed name or attribute of an element both versions hold keeps its own check. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): read an attribute called name or attributes as an attribute when the draft keeps its check The owner of `.../attributes/name` was taken as `.../attributes`, so a changed `name` attribute inside a paragraph with a text check was dropped again. The owner is now read by segment, attributes first. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): draft text checks from text in runs added whole, and name a text element by its own id before its container Text typed into a run the diff lists whole was never read, so the draft kept only a presence check on the run, which wrong text also passed. Such text now marks the nearest element both ends hold, or, when none can take the check, the added element's own text. The climb also names an element by an identifier of its own (id, xml:id, xid) before a larger container, so a headline under a long layer keeps a whole-text check that survives re-split runs. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(desktop): expect the whole-text check where #8988's pairing test meets text checks On this branch a changed element's text is drafted as `text_equals` on the element, so the identity pairing test now expects that check, under the same `[@role=body]` key. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(skills): build a computer-app environment from a request The product skill gains references/desktop-build.md, a procedure that takes Claude from a request such as "an Inkscape environment where the agent changes a poster's headline" to a witnessed bundle and a report, through start_build_computer, act_on_build_computer, the build capture tools, export_build_environment and check_build_task: questions asked once upfront, pauses only when blocked, a seed made in the app, content checks by selector, a witness with a near miss and a reset in a fresh computer, and solver runs only within a budget. setup installs the new reference; skillVersion 61. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(skills): fold the blind Inkscape rebuild's friction into the build skill key takes an array of one chord's keys; check for a text cursor before typing in apps with single-key shortcuts; in Save As, click Documents and type only the file name; anything saved before seed, the Trash included, is part of the seed; dismiss first-run dialogs before seed; a bundle's app starts with the seed's files but need not open the document; Inkscape's window class is Inkscape. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(skills): fold the held-out AbiWord build's friction into the build skill Name the act fields and link the guide; a region screenshot becomes the latest frame; a drawn caret is not focus; confirm dialogs with Return; check the dialog's folder when the app shows only a name; an XML file the draft shows as lines was read as text, so use short line checks; and keep text checks to the value the edit types, since some apps save a reloaded document differently. Later witness stages take no directory. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(skills): state the export's seed rule, the Trash exception and draft refusals exactly The export refuses a task unless at least one check fails on the seed and every check passes at the end; a file moved to the Trash is left out of the seed, so trash a stray file rather than starting over; a null draft names its cause in draft_refusal; and a line check on a format read as text only approximates a value, so bound it and call it approximate. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(skills): teach the build skill to use DOCTYPE and whole-text checks The skill now checks text an app splits into runs whole with text_equals or text_contains, finds an AbiWord paragraph by its xid, and drops the Writer revision-id workaround. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * docs(skills): point text checks at the paragraph, and say what the scrub replaces From the AbiWord rebuild: point a text check at the paragraph or another element the edit keeps, never at a run a reopened document splits; how to name an unchanged element; the export replaces the host name, user@host and home path while a bare user name stays; AbiWord's entry and class. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): keep every text check on a cut element undecided, in both evaluators The XML reader drops text deeper than its depth limit but keeps later siblings, and an element's whole text joins both sides, so a text_contains on a cut element could match text the file never holds ("at 3" + deep "WRONG" + ":30 PM" read as "at 3:30 PM"). Any text check on an element of a cut view is now undecided again, as #9050 had it, in REA's evaluator and the exported verifier; a text node or attribute value stays decided. The text check tests assert facts rather than sentences. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * feat(desktop): open the seed's document at launch, and say when a seed was made for the task An exported app now opens the seed's one document at launch when its desktop entry's Exec takes files, so a solving agent starts in the document rather than on the app's start screen; a command that already names a file is kept, and several documents or an Exec without a file code keep the launch and say why. `--synthetic-seed`, which the build export passes, words the README and limitations for a seed made in the build computer instead of asking to keep the user's data local. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): count only a document the launch names, absolute or from its working directory, as already opened A settings path under the home, such as geany's --config, kept the seed's document from opening, and a relative document argument was missed and appended again. An argument now counts only when it resolves, from the captured working directory, to a file the capture gave the document role. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(desktop): check the seed and launch wording against their formatters, and the facts on their own The tests froze the README's and limitations' sentences, so a harmless rewrite broke them. The wording now comes from exported formatters (`seedWording`, `seedLimitation`, `launchDocumentNote`, `NO_LAUNCH_DOCUMENT`), the tests compare against those, and check the facts separately: which document opens, the respelled path in the command, the README naming it, `--synthetic-seed` and the Evidence's `synthetic_seed`. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * refactor(desktop): plan the export's seeds and launch in their own module The merged export module passed the file and function size limits; the seed digests and the launch command now come from `DesktopExportLaunch.ts`. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): keep a reset applied when the screenshot asked after it fails Once data/reset and the ready wait succeed, act_on_build_computer's reset is applied in the action log and returns success; a screenshot asked of it that is then refused or throws is named in the result's screenshot_error, with no frame, so the next after_reset stage is not refused. Cancellation still propagates. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): keep checks for elements without text beside a text edit, and normalize DOCTYPE system id line endings A review found two gaps. - A paragraph's whole-text check covered every element added or removed inside it, including ones holding no text: an image frame, a line break. A solver who typed the text but never inserted the image passed. Such an element now keeps its own presence or absence check. - The Python reader kept a DOCTYPE system literal's CRLF where TypeScript normalizes it to LF. Both now report LF. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): wait out long build commands, fail a reset that restores nothing, and tie the witness to its bundle A review found five gaps. - Commands in the computer answer only when they end, but fetch stops waiting for headers after 300 s. So a slow export, draft or start always failed. A command with its own deadline now waits on node:http, bounded by that deadline. - after_reset held whenever some check failed, so a reset that changed nothing passed on the near-miss state. Each check must now come out as it did at the seed. - A failed or partial apt-get update left lists that looked complete, so a second start skipped the update. The update is now recorded only when it succeeds, and rerun otherwise. - A witness on a bundle computer graded whatever task it was given. Its seed stage now refuses a task whose steps or checks differ from the bundle's. - A task too large for one command argument (96 KiB) is refused before the export runs, instead of failing with E2BIG in the computer. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): advertise a checkpoint-name pattern that compiles in every JS regex mode main's advertised-pattern test compiles every input schema pattern in v mode, which refuses an unescaped hyphen in a character class. The checkpoint name's hyphen is now escaped, and it means the same in u mode. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * test(desktop): import textSyntaxFormat once in the state format tests main's mixed-case JSON suffix fix and this branch both added the import, and the merge kept both. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> * fix(desktop): keep a completed reset, an unread seed, and a failed apt update A cancel after data/reset no longer drops the restore from the action log, so a near miss cannot treat the restored seed as an edit. An unreadable seed check that later passes leaves after_reset undecided instead of failing the witness. apt-get update is kept only when Error-Mode reports success and a package list was fetched. The controller reply schema lives with the controller, and capture progress lines are read by their domain parser. * fix(desktop): keep a whitespace-only run under the paragraph text check A run whose only text is whitespace is still text the paragraph check reads. It no longer gets its own presence check, which a later save of the same words without that run would fail. Images and breaks, whose text is empty, still keep their own checks. * fix(desktop): keep an unread reset and a cut text check undecided after_reset stays undecided when any check was not read, even if another check still fails and the unread result's boolean matches the seed. A text check for a child past a truncated view is undecided in both evaluators. A reset whose reply is lost blocks a near miss until a later input or a reset whose reply arrives. * fix(desktop): retry a witness recording and a published record's cleanup A recording that fails to end or download is not written into the witness folder, and a manifest already received is fetched again without ending the recording twice. A folder that is published while its handles fail to close keeps that record and retries the same cleanup instead of writing the folder again. * fix(desktop): keep a published witness error and tag an app data refusal A second failure to close a published witness returns the original error, so its record and cleanup lease stay bound. A reset the app itself refuses is an AppDataRefusal; only a reply that is not that refusal stays unsettled. A missing child of a cut view is undecided with the same description in both evaluators, and the shared cases cover it. * fix(desktop): align a cut view's missing text with the Python check A missing child of a truncated view is undecided for the same sentence in both evaluators, including the pointer the view does not contain. * fix(desktop): keep one published witness cleanup across retries A published record and its bound cleanup error stay together. A later close failure returns that error, and the unused path that built a fresh adapter error is gone. * fix(desktop): keep a bundle task read inside the witness result A cancelled, refused, or failed read of the bundle's task files returns the witness's Result, including a cleanup lease when a handle would not close. * fix(desktop): check build runtime export capability * fix(desktop): provide complete runtime repack command * fix(desktop): retain runtime probe failures * fix(desktop): recognize captured file URL documents * fix(desktop): render captured file URLs at launch * fix(desktop): isolate launcher URL encoding * test(desktop): move launch execution to boundary suite * style: fix desktop launch test formatting * style: format native command path schema --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Cursor Agent <cursoragent@cursor.com> Co-authored-by: morluto <morluto@users.noreply.github.com> Co-authored-by: morluto <76467478+morluto@users.noreply.github.com>
English · 简体中文 · 繁體中文 · 日本語 · 한국어 · Türkçe · Русский · Tiếng Việt · ไทย · Deutsch · Español · Français · Українська · Polski · Português (Brasil) · العربية · فارسی · Bahasa Indonesia · Italiano · Ελληνικά · മലയാളം · Roman Urdu · Română
REA: Reverse Engineer Anything
One MCP for reverse engineering across binaries, applications, and runtime behavior.
See a feature you like. Understand how it works, down to the binary level.
Quick start · How REA works · What you can analyze · Showcases · Build RL environments · FAQ · Documentation
npx rea-agents setup
Important
Get REA from morluto/rea, rea.tools, or the rea-agents npm package. Follow the installation and release links from these official sources. A similar repository name or copied README does not establish that a download is official.
See a feature in an app that you want in your own product? Ask your agent to investigate it with REA. It can inspect the app without its source code, explain how the feature works, show the evidence, and build a version for your project.
REA connects your agent to tools for inspecting native binaries, JavaScript and Electron apps, .NET assemblies, and websites. You can also use the same tools from your terminal. Analysis runs locally, and results include the evidence and limitations behind each conclusion.
Setup registers REA with your agent and installs matching workflow instructions. Native analysis can use an existing Hopper, Ghidra, IDA or Kuna installation; setup can optionally install Hopper with approval. Static JavaScript analysis needs no native analysis engine.
Visit the REA website for setup instructions, illustrated guides, and real case studies.
Quick start
On Linux, Ghidra, JADX, firmware tool launches and static JavaScript/Electron
analysis require a procps-compatible ps on REA's PATH for process ownership
checks. Recovery steps are in docs/getting-started/installation.md under
"Linux owned-process prerequisite".
Set up your agent
With Node.js and npm installed, run:
npx rea-agents setup
Choose your agents, review the proposed changes, and approve them. Setup adds REA's MCP server and matching workflow instructions, with backups of existing configuration. Restart your agent afterward. Setup records the selected runtime and REA entry script directly. If acquired through a package runner, keep its cache entry available. Run package acquisition from a trusted directory outside the target project. Installation and recovery details are in the setup guide below.
Setup supports Claude Code, Codex, Cursor, Gemini CLI, Grok Build and other agents. See installation and setup for provider configuration and manual MCP registration and Claude Code/Codex plugin installation.
MCP uses compact input schemas by default to reduce context overhead. Set
REA_MCP_INPUT_SCHEMA_PROFILE=full in your MCP registration when you need the
complete advertised field guidance and nested validation details. Both profiles
keep the full tool set and canonical server validation. See
"Compact input schema profile" in the MCP contracts guide.
Complete output schemas are advertised by default. If your client caps the
complete tools/list frame, set REA_MCP_OUTPUT_SCHEMA_PROFILE=compact to
bound each advertised output schema, or off to omit output schemas entirely.
Both settings keep canonical server-side validation and complete structured
results. See "Output schemas" in the MCP contracts guide.
Ask your agent
Understand how search works in the Notes app, show me the evidence, and build a
similar feature for my project.
Replace Notes with your target app and the feature you want to understand.
Use the terminal
Inspect an extracted JavaScript/Electron app directory or ASAR:
npx -y rea-agents@latest analyze-javascript-application /absolute/path/to/app --json
The result includes modules, imports, Electron boundaries and their evidence.
Replace the path with your target, such as "D:/apps/example" on Windows.
To install the rea command for regular use:
npm install --global rea-agents
rea --help
For native analysis, configure a provider first. See the CLI and Evidence guide for native commands, provider selection, snapshots and scripting.
Update REA
REA changes quickly, and new releases include frequent bug fixes. Keep your installation up to date.
For a CLI installed with npm install --global or bun add --global:
rea update
rea --update is an alias for this verified workflow and accepts the same global
output flags. Use rea update when combining it with other command controls.
Updater commands retain at most 4 MiB of stdout and stderr combined. Oversized metadata replies are rejected after the command exits; verbose installs continue with bounded diagnostics. Human progress output respects terminal backpressure and waits for submitted writes to finish.
Bun global updates require Bun 1.2.17 or later. With an older Bun release,
upgrade Bun manually before running rea update.
To refresh your agent registrations and skill, run the setup command printed by the update.
If you use npx, update your agent setup with:
npx rea-agents@latest setup
Review the setup changes and restart your agent. For one-off CLI commands,
use npx rea-agents@latest followed by the command. With Bun, use
bunx rea-agents@latest in place of npx rea-agents@latest.
How REA works
Your agent calls REA through MCP to inspect the target and trace relevant code. REA returns findings with their evidence. The agent uses them to ask follow-up questions, explain the behavior, or write and test an implementation. CLI commands use the same workflows.
What you can analyze
REA requires Node.js 22.x (>=22.23.4), 24.x (>=24.22.1), or 26.12.1+, plus npm. Additional tools and host support depend on the target:
| Target | What REA returns | Requirements and guide |
|---|---|---|
| Desktop installations | Runtime candidates, supporting files, component markers and next probes | Built-in bounded directory/header inspection; desktop applications |
| Native binaries | Pseudocode, assembly, strings, symbols, calls and references | Hopper, Ghidra, IDA or Kuna; native analysis |
| Offline ELF layout | Sections, segments, original symbols/relocations and static mitigation candidates | Caller-supplied pwntools on Linux x64; binary diagnostics |
| Go binaries | Embedded compiler version, modules, replacements and build settings | Built-in static ELF/PE/thin Mach-O inspection; no Go installation required |
| GPU code objects | CUDA/HIP targets, selected kernel parameters and resources, retained PTX text | Built-in static ELF/PE/fatbin/CUBIN/PTX inspection; GPU code objects |
| Model artifacts | GGUF / safetensors metadata, tensor declarations and derived totals | Built-in header inspection; optional full-file hash; model files |
| IL2CPP metadata | Unity global-metadata.dat assembly images, type and method names and index ranges | Built-in static decoding of versions 24, 27, 29, 31 and 39; no Unity required |
| Godot PCK | Archive inventory and extraction, original header/directory facts and selected content identities | Built-in static formats 2, 3 and 4; no Godot required |
| Rust binaries | Embedded cargo-auditable crate names, versions, sources and dependency edges | Built-in static ELF/PE/thin Mach-O inspection; Rust binaries |
| PE certificates | Authenticode WIN_CERTIFICATE entries and PKCS#7 signer/X.509 identity fields | Built-in static PE inspection; PE certificates |
| PE imports | Import and delay-load DLL names, ordinals, IAT RVAs and pefile import hash | Built-in static PE inspection; PE imports |
| ELF notes | PT_NOTE/SHT_NOTE owners, types, GNU build-id and ABI-tag fields | Built-in static ELF inspection; ELF notes |
| Binary hardening | PIE, NX, RELRO, CFG and related header flags for ELF, PE and Mach-O | Built-in static header inspection; binary hardening |
| Apple IPS crashes | .ips metadata, threads, images, registers and documented crash classes | Built-in static JSON inspection; Apple IPS |
| Linux kernel modules | .modinfo keys, __versions CRCs, this_module presence, and signature trailer | Built-in static ELF inspection; kernel modules |
| Erlang BEAM | IFF chunks, atoms, exports, imports, lambdas and the Code header | Built-in static BEAM inspection; Erlang BEAM |
| WASM artifact | Validation, sections, exact import/export WAT forms and decoded WAT | Caller-supplied WABT 1.0.42; inspect_wasm_artifact |
| EVM bytecode | Dispatch selectors, byte offsets, inferred arguments and mutability | Local raw/hex carrier on Linux x64; offline EVM guide |
| SQLite databases | Tables, columns, indexes, views, triggers and selected ordinary-table rows | Local database/WAL snapshot; Node.js 24.x >=24.15 or 26+ with native SQLite limits |
| Recorded Linux crashes | Raw notes, every recorded thread's registers/signals and optional mapping candidates | Caller-supplied pwntools on Linux x64; optional GDB/pwndbg; recorded crashes |
| JavaScript / Electron | Modules, imports, source maps, routes, IPC and native add-on relationships | Node.js and npm; Linux: procps-compatible ps; application analysis |
| Websites | Page structure, scripts, network observations and requested screenshots | A Chrome-family browser; browser analysis |
| Saved network captures | Requests, responses, exposed payloads and source locations | HAR or Fiddler Everywhere SAZ; mitmdump on Linux for native mitmproxy; capture guide |
| .NET assemblies | Metadata, CIL instructions, declared native dependencies and build comparisons | None (static only); managed-code guide |
| Android APKs | Manifest declarations, classes, decompiled methods and references | Headless JADX and a full JDK on Linux/macOS/Windows x64; Android guide |
| APK signatures | v1/v2/v3/v3.1 signers, certificate fingerprints, key rotation and digest self-consistency | None (static only); inspect_apk_signature |
| Android devices | Devices, packages, processes, logs, dumps, screen captures, and APK/file transfers | Caller-supplied adb (emulator or hardware); ADB guide |
| JEB projects | Engine-backed project units, artifact digests and type/method pseudo-code | A caller-started JEB client serving MCP, selected with REA_JEB_MCP_URL; JEB guide |
| Android resources | Decoded manifest, string tables, locales and version facts | Caller-supplied Apktool launcher; resource guide |
| Flutter apps | Dart snapshot hash, per-ABI library digests and engine build provenance | None (static APK parsing); Flutter guide |
| Firmware | Regions, extraction results and native-analysis handoffs | Binwalk / Unblob on Linux; firmware guide |
| Recorded packets | Selected fields, repeated values, filters and decode-as rules | Linux: TShark 4.x + prlimit; capture guide |
| Packet summaries | Endpoints, DNS, TLS SNI, protocol distribution and HTTP requests | Caller-supplied tshark and capinfos (Linux-verified); pcap guide |
| Recorded logic signals | Decoder annotations, sample ranges and diagnostics | Linux: sigrok-cli 0.7.2 + prlimit; signal recordings |
| Packages and resources | File inventories, digests, plists, Apple bundle anatomy and extracted resources | Artifact and JavaScript guide, Apple applications |
| Process behavior | Terminal output, interactions, exit and filesystem observations, and run comparisons | Linux/macOS with a native PTY; process capture |
| PyInstaller executables | CArchive members, Python version, entry scripts and extracted bytecode | None (static only); Python packaged applications |
Static JavaScript and .NET inspection read the supplied files without running the application. Runtime capture runs or interacts with the selected target using your user permissions; each runtime guide describes its effects.
For offline native-binary inspection, REA also supports caller-supplied GNU objdump and Rizin. See the reverse-engineering tools guide for provider setup, CLI commands, and Evidence limitations. REA does not install or update either provider.
Native formats and host support vary by provider. See
Hopper and Ghidra setup, the
IDA guide (including an optional native MCP bridge), the
Kuna decompiler guide, and
experimental Windows Ghidra support.
Ghidra also supports 16-bit DOS analysis.
For large binaries, raise the provider's startup deadline with
REA_GHIDRA_STARTUP_TIMEOUT_MS or REA_HOPPER_STARTUP_TIMEOUT_MS.
For provider selection, see the CLI guide.
Check release availability
for features added since the latest npm release.
Showcases
DX-Ball: reconstruct a sound-pan calculation
Follow a sound call into its position-to-pan helper, inspect the instructions, and turn incomplete pseudocode into C. The reconstruction passes 3,205 original-x86 cases and reproduces all 63 compiled function bytes.
Read the case study · Reconstruction repository
Notion: trace the Electron clipboard bridge
Find the renderer's clipboard API, follow it through preload and IPC into the main process, and inspect the rich clipboard format.
TH04: recover a DOS bullet-ring calculation
Inspect the original PC-98 game's 16-bit instructions, recover the fixed and aimed angle calculations, and compare the reconstructed C++ with the historical compiler output.
Read the case study · Reconstruction repository
Build RL environments with REA
Record a web app session, turn its API into tools, and export an environment an AI agent can use. Run it with AgentEnv, then add tasks and reward checks for RL training or evaluation.
Read the guide · Watch the video
If you've used REA on something interesting, we'd love to see it. Share your case in an issue or a pull request, including the target, your question, how REA helped, and what you found.
Build RL environments
Give an AI agent a local environment to practice tasks such as finding an issue or creating a folder, without running the original app. Use it for reinforcement learning (RL) or to test how well an agent completes tasks.
REA turns recorded API requests and responses into tools the agent can call. The environment uses the recorded data and supports changes only when the recording confirms their effects.
With REA set up in your agent, replace the example URL and folder tasks below with your own:
Use REA to build an RL environment from my test web app at
http://127.0.0.1:3000.
Cover listing, creating, renaming, and deleting folders. You may create and
remove test folders. Leave existing data unchanged.
Save it in a folder named my-env. Check which operations work and tell me
how to run it.
Use the command line instead
Use these commands if you already have a file containing the app’s recorded browser requests and responses. Running the result also requires AgentEnv and Docker.
rea environment-build ./session.har ./my-env --name my-env --json
agent-env run ./my-env
REA exports a runnable environment with a basic test to check that it starts and responds. Use it with your RL framework to define task goals, add reward checks, and run training.
Build an environment · Environment reference · Watch the demo on X
FAQ
Which agents can use REA?
Any agent that supports local MCP servers. Setup configures the supported agents; other clients can use manual MCP registration.
Do I need a native analysis engine?
Deep native analysis uses Hopper, Ghidra, IDA or Kuna, depending on the target and operation. Static JavaScript and .NET inspection work without a native analysis engine. Setup can install Hopper after approval; Ghidra, IDA and Kuna use your existing installations. See provider setup for the options and their supported targets.
Do I need to start Hopper first?
REA starts Hopper when an operation needs it. On macOS, a first-run dialog may ask you to choose demo mode or activate your license. Complete that prompt in the same user session before your first CLI/MCP operation or unattended job. Open Hopper manually if needed. See Hopper startup and troubleshooting.
What does installing the skill from skills.sh do?
The skill supplies investigation instructions for your agent. Use npx rea-agents setup
to register REA's MCP server and install the matching instructions, then restart
your agent. See skill-only installation.
What code does REA return?
Native analysis returns pseudocode and assembly. JavaScript/Electron analysis recovers modules and their relationships. Your agent uses these findings to write and test an implementation; the showcases give worked examples.
Does REA upload my app?
REA analyzes targets locally. Your agent receives the tool results, and its model provider has its own data policy.
What should I do if I hit a bug?
Update first; a recent release may already fix it.
For a CLI installed with npm install --global or bun add --global:
rea update
For agent setup through npx:
npx rea-agents@latest setup
If you're using an agent, complete the setup refresh and restart it. Retry the same task. If the problem persists, open an issue with your REA version, target type, steps to reproduce and error output.
Documentation
Start with the website's worked guides. For exact options, prerequisites and result contracts:
- Installation and setup: agent registration, provider configuration, updates and uninstall.
- Readiness and troubleshooting: diagnose one agent or analysis engine.
- CLI and Evidence: commands, provider selection, snapshots, import/export and exit statuses.
- MCP contracts and agent prompts: tool results, sessions and guided investigations.
- Tool catalog: generated inventory of tools, providers and CLI commands.
- Roadmap: planned work and capability trackers.
Report vulnerabilities through SECURITY.md.
Star history
🎉 400,000 GitHub stars — thank you!
Thanks to everyone using REA, reporting bugs, requesting features, testing builds, and contributing fixes.
Disclaimer
See DISCLAIMER.md for responsible-use guidance.
The project's position on illegal or unauthorized use covers all REA versions, distribution channels, and platforms; each copy retains its applicable licenses.
REA is for lawful research, education, interoperability, software preservation, and analysis of software you own or are authorized to examine.
Contributing
We'd love your help with REA! Open an issue to report a bug or suggest a feature, or send a pull request to improve the code or docs.
See CONTRIBUTING.md for development setup and checks, testing for verification lanes, and the architecture map for selected layers and dependency paths.
License
REA-authored code is licensed under MIT. Third-party material retains its own licenses and copyright notices; see Third-party notices for the dependency inventory, preserved license texts, and external-tool terms.
You must legally obtain and be authorized to use commercial tools, including IDA Pro / Hex-Rays decompilers, Hopper, and JEB, under their applicable licenses. REA grants no tool license, key, or activation entitlement. Trials and demos remain subject to vendor terms, including user, feature, automation, and output sharing restrictions. Cracked copies, stolen keys, unauthorized key sharing, and bypassing license restrictions are not authorized. This guidance does not amend MIT or third-party licenses.