mirror of
https://github.com/newliver666/apk-reverse.git
synced 2026-09-28 21:56:59 +02:00
inliver233
2beade67f2
An external audit of this repository landed on four structural findings, and this pass is the
response to them. Everything below was re-measured on this tree before being acted on; two of the
audit's claims needed a correction of my own, and both corrections are noted in place.
**The gates were optimistic, and one of them did not read content at all.**
doctor.py the capability table was hard-coded (`repack + sign = bool(java) and python`), so a
machine with a JDK and no zipalign/apksigner was told it could re-sign; the script
inventory was a hand-written list of 28 against 51 files on disk, omitting the nine
most important scripts. It now scans the directory (`registered/checked = 52 / 52`,
unjudgeable entries counted as `unknown`, never skipped) and resolves capability as a
real requires-closure, reporting BLOCKED with the missing piece and an installable
next step. Measured: the old doctor reports `[OK] repack + sign` on a JDK-only PATH
where keytool, jarsigner, zipalign, apksigner and all eight smali jars are absent.
The report is allowed to look bad; that is the point.
check_commands.py (new)
anchor checks prove a section exists, never that a command runs. This parses each
script's argparse table with `ast` and checks every documented invocation against it.
146 commands / 41 interfaces, 0 drift. It caught `sig_probe.py --live --pkg`, which
does not exist -- the reference had been wrong since it was written.
check_routing.py (new)
the symptom index now has one home (`SKILL.md`, because symptom -> file must stay one
hop) and the inventory has another (`references/routing.md`). This gate proves the
mirror agrees, that every reference file is named, and that every script is named.
**so_constpatch.py was corrupting APKs it rebuilt.** The writer copied filename and timestamp onto a
fresh ZipInfo and recompressed the whole archive: 435-entry real package, `extra` lost on 205 entries,
`external_attr` rewritten on all 435, 132 untouched entries recompressed, and `resources.arsc` pushed
to an unaligned offset. Android's own `zipalign -c -v 4` fails the old output and passes the new one;
a 26-case fixture suite pins it, and `extractNativeLibs="false"` is now refused by default with a
pointer at `repack.py`. Fixing it exposed the same defect in `repack.py`, where `^lib/[^/]+\.so$`
cannot match `lib/arm64-v8a/*.so` -- so its alignment handling had never once fired on a real APK.
**The package was not self-contained.** 59 references pointed at the repository root, which `npx
skills add` does not install. They now converge on `references/evidence-summary.md` (capability,
conclusion, strength, evidence that ships) plus `evidence/` with the machine-readable companions: 8
references remain and each is explicitly labelled "repository root, not shipped". A drill in which
the new files were copied to an isolated directory and asked three questions answered all three.
**SKILL.md: 572 lines / 15,733 tokens -> 398 lines / ~10,900 tokens**, without losing the two things
that had to stay. The symptom index remains inline; the reference and script indexes moved to the
on-demand routing file. The freed budget went into making the rules *stronger*, not shorter: the
constraints section is now a table of thirteen non-negotiable constraints each paired with the check
that catches a violation, and the tooling obligations are instructions whose violation is a defect
(hedging language in SKILL.md: 1 line, down from a paragraph of hedged advice). Long references gained
a navigation head so a reader can decide from the top whether to load one.
**New: an executable test layer.** `tests/` (unit, CLI contract, no-device integration) plus
`pytest.ini`, `tests/run_tests.py`, a minimal CI job, and `evals/evals.json` written but never run --
stated as such in the file. 608 passed / 158 skipped / 25 xfailed. The 25 xfails are a work order, not
a pass: thirteen scripts report a missing input by exiting 1 instead of 2, seven let a traceback
escape, four treat `--help` as an error, and `preflight --json` lacks a `status` field. A regression
row for `dex_patch_bytes.py` reproduces a previous artifact byte-for-byte
(`67ae1e58e7a44777e5ac8f9f823ab21e…`) and additionally pins the four bytes that may move.
**Two real defects the new tests found in the tooling itself.** `check_repo.py` decoded subprocess
output with the locale encoding, so on a GBK console a script printing an em dash in its usage line
was reported as crashed; it now decodes UTF-8 explicitly. And a CLI contract test drove a *real
attached device* through `coldstart.py` -- logcat, screenshots, five orphan `adb logcat` processes --
which is fixed twice over: child processes run in a private temporary cwd, and the fifteen
device-bound scripts are never executed beyond `--help`/argument parsing.
**Quoting on the device side is now one implementation.** `devsh.su` escaped only double quotes, so
`;`, `$( )` and backticks in a caller-supplied value still reached the device shell while
`coldstart.Device.shell` quoted correctly. Both use `scripts/device_shell.py` now: POSIX single-quote
escaping, argv arrays by default, and identifier validation that refuses a malformed package, path or
component with exit 2 instead of a traceback. The integration test that used to pin the *defect* (it
asserted `su -c "…"` with `\"` escapes) now asserts the property, across five metacharacter payloads.
Gates: check_repo 0 problems (including the content-reading identity scan over git ls-files),
check_refs 0 dangling / 0 warning, check_routing coherent (53 symptom / 45 reference / 54 script rows),
check_budget 0 failure / 0 warning with the always-loaded body now measured in tokens rather than in
"narrative lines with the index excluded", check_commands 146 checked / 0 drift, pytest 608 passed.
Not done, and stated rather than implied: CI has never run on a runner; the evals have never been
executed; the 25 xfails are open; no device-side install/launch was performed in this pass.
tests/ - the executable half of this repository's evidence
tests/benchmark.md is a record: what happened when a route was run once, by hand,
against a real sample and a real device. This directory is the other half: checks that run
on every commit with nothing but the standard library and pytest.
The two are not substitutes. A benchmark row can establish that a route works on a hardened target; it cannot tell you that yesterday's refactor broke the dex width table. A test in here can tell you that every time, and it can never tell you whether the route achieves the goal on a real app.
python -m pytest -q # whole suite
python tests/run_tests.py # same thing, with a clear message if pytest is missing
python -m pytest -q -m unit # one marker
python -m pytest -q -m cli # exit codes, flags, --json
python -m pytest -q -m integration # no-device flows
python -m pytest -q -ra # list every skip and xfail reason
What the markers mean
unit- pure functions and parsers of real shipped scripts: the dex width table and branch arithmetic, the scheme-less protobuf decoder, the blob rotation search, the leak rules, and the establisheddex_patch_bytesregression hash.cli- one subprocess run per script per invocation shape:--help, an unknown flag, no arguments, and every input pointed at a file that does not exist.integration- flows that were made of several parts: adb argv construction under a fake runner, logcat signal counting, tombstone parsing, the two-sample launch-health verdict.
Nothing here needs a device
No test in this suite opens a frida session or touches the network. The device-dependent behaviour is tested by intercepting the command that would have been run and asserting on the argv list.
The exception is deliberate and enforced: contract.DEVICE_OR_NETWORK_BOUND names the 15
scripts that reach a device once their arguments parse, and those are never executed with
complete arguments — only --help and an unknown flag, which argparse answers before any
device call. That blocklist exists because a development host may well have a phone attached
(the machine this harness was written on did), and an earlier version of the CLI contract test
drove it for real. General safety net: every child runs in a private temporary working
directory, so a script that defaults an output path relative to cwd cannot write into the
shipped tree.
Anything that genuinely cannot be reached this way is not faked into a pass — it is listed as a
coverage gap in docs/tool-verification/EXTENSION-test-harness.md.
Skipped is not passed
Two kinds of non-passing outcome are kept separate on purpose:
skip- the check does not apply here.tools/_work/bench/**is git-ignored, so thedex_patch_bytesregression hash self-skips on a fresh checkout and in CI, printing the reason. A --json script that needs a sample to reach its JSON path skips too.xfail- the script is present and does not meet the repository contract. The reason carries the measured value, for examplecontract says exit 2, measured 1. These are work items, not flaky tests.
pytest -q -ra prints both lists. The xfail list for the CLI layer is the exit-code
unification work order; it is reproduced in EXTENSION-test-harness.md with owners.
Ground rules
- This suite never edits anything under
skills/. A defect found here is registered in the harness report. Fixing a script's exit code or quoting strategy is its owner's change, and doing it from a test would hide the drift the test exists to expose. - The bench tree is evidence and is read-only here. The regression test copies its input
into
tmp_pathbefore patching, sotools/_work/bench/l1l3/dex/classes.patched.dexcannot be overwritten by a test run. - No real target identity is committed. The leak-scanner decoys are synthesised in
tmp_pathat run time, and token-shaped decoys are assembled from fragments so that this directory cannot itself become a finding. - Fixtures in
tests/fixtures/are small, synthetic sample files.tests/fixtures/apk/**is owned by a different work item and is not touched from here.