Files
stack/packages/mosaic/framework/tools/git/test-wrapper-guard.sh
T
Hermes Agent 20d86e392b
ci/woodpecker/pr/ci Pipeline was successful
fix(guard): end the home match at a shell word boundary, not at whitespace
Rounds 8 and 9 of the same class, in the two halves of one line.

The path arm required the home token to be followed by `/`. That silently
made `$HOME` itself -- the exact target the rule names -- legal: `git worktree
add $HOME` cleared a guard whose message is "this checks a repository out
under $HOME". Reachability is not theoretical; the command succeeds against an
empty home directory. Trailing `/` was then admitted, and with it every
terminator that is not whitespace: `$HOME;`, `$HOME&&`, `$HOME|`, `$HOME&`
and end-of-string all cleared, 25 shapes in all.

The fix that did not happen is worth recording, because it was mine. The brief
for this round prescribed a closed continuation class, `([^A-Za-z0-9_.-]|$)`,
on the reasoning that terminator sets are open and continuation sets are
closed. That is true of some axes and false of this one: `+ @ , : = %` all
continue a FILENAME, so `$HOME+bak/wt` and five siblings like it would have
been refused -- a new over-block traded for a closed bypass, which is not a
trade. The implementer measured the six counterexamples and declined the brief
rather than pick between two acceptance conditions that cannot both hold. They
are now permanent fixtures; a rejected over-block that nothing pins comes back.

The axis that IS closed is word termination, and it is closed by specification
rather than by anyone's imagination: POSIX fixes the unquoted metacharacter set
at space, tab, newline, and | & ; ( ) < >. So the path normalizer marks those
as an internal word boundary, in the same state machine and by the same
mechanism as the existing literal-dollar and literal-tilde markers, which is
what lets a QUOTED or escaped metacharacter stay word content: `"$HOME;bak"`
is one word and must be allowed. A raw marker byte arriving in the input is
encoded first, so input cannot forge or suppress a boundary. The home token
must now be preceded by start, `=`, or a boundary, and followed by a boundary,
`/` for a descendant, or end.

Verified by oracle rather than against the brief -- `bash -c "printf '%s' WORD"`
performs expansion and quote removal without executing, so the expected verdict
comes from the shell instead of from the reading that has now been wrong once.
Fixtures 198 -> 230; the new ones are red at both prior heads (15 failing at
4b8eba95, 21 at 3d0a882a), so they measure the change rather than passing on it.

Known and deliberately not addressed here: a checkout target that never names
$HOME at all. A relative target resolves against the cwd, and every agent seat
on this host runs with a cwd under $HOME, so `git clone URL` with no target at
all lands in $HOME and is invisible to a rule that matches home spellings.
That is a different rule -- it needs the effective cwd, which `cd` inside the
command can move -- and it is filed separately rather than becoming round ten
in this file.
2026-08-13 05:08:01 -05:00

606 lines
55 KiB
Bash
Executable File

#!/usr/bin/env bash
# test-wrapper-guard.sh — hermetic behavioural regression for wrapper-guard.sh.
#
# Resolves no credentials, touches no network, and creates no repository: the
# guard reads a hook payload on stdin and answers with an exit code, so the whole
# contract is testable from fixtures.
#
# The fixtures are written to a temp file rather than passed inline, and this is
# not stylistic. The guard inspects the literal text of the Bash command it is
# handed. A test that embeds `git clone ... $HOME` inside its own command line
# trips the guard on the harness instead of on the fixture — which is exactly
# what happened the first time this was checked by hand. Substring matching over
# whole command text is the guard's deliberate fail-closed posture; a test that
# does not account for it silently measures the wrong thing.
#
# Exit: 0 = every fixture behaved as specified · 1 = at least one did not
set -uo pipefail
HERE="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd)"
GUARD="${1:-$HERE/wrapper-guard.sh}"
[ -x "$GUARD" ] || { printf 'test-wrapper-guard: not executable: %s\n' "$GUARD" >&2; exit 2; }
TMP="$(mktemp -d)"; trap 'rm -rf "$TMP"' EXIT
FIXTURES="$TMP/fixtures.tsv"
# Each line: <expected-exit> TAB <hook payload> TAB <what it proves>
# [ TAB <substring the block message must contain> ]
# 0 = allowed, 2 = blocked. The optional fourth field is how the remediation
# itself gets checked; without it a block is only asserted to have happened,
# not to have been useful.
{
printf '2\t{"tool_input":{"command":"git clone https://example.invalid/x ~/wt"}}\tcheckout into $HOME is refused\n'
printf '2\t{"tool_input":{"command":"git worktree add ~/wt topic"}}\tworktree into $HOME is refused\n'
printf '2\t{"tool_input":{"command":"g\\"it\\" clone https://example.invalid/x $HOME/wt"}}\ta double quote inside git does not hide a checkout\n'
printf '2\t{"tool_input":{"command":"g'"'"'it'"'"' clone https://example.invalid/x $HOME/wt"}}\ta single quote inside git does not hide a checkout\n'
printf '2\t{"tool_input":{"command":"g\\\\it clone https://example.invalid/x $HOME/wt"}}\tan unquoted escape inside git does not hide a checkout\n'
# Path words use the same quote/escape state machine as names, but preserve
# substitutions so HOME remains visible. Quotes do not split the path word.
printf '2\t{"tool_input":{"command":"git clone x \\"$HOME\\"/wt"}}\ta closing quote between HOME and slash does not hide the path\n'
printf '2\t{"tool_input":{"command":"git clone x ${HOME}/wt"}}\tthe braced HOME spelling is the same home path\n'
printf '2\t{"tool_input":{"command":"git clone x \\"${HOME}\\"/wt"}}\tbraced HOME may also end a quoted span before the slash\n'
# The target may be HOME itself. End-of-command and whitespace terminate the
# token just as a slash does; punctuation that can extend a path does not.
printf '2\t{"tool_input":{"command":"git clone x $HOME"}}\tthe unbraced variable may name HOME exactly\n'
printf '2\t{"tool_input":{"command":"git clone x \\"$HOME\\""}}\tquotes do not change the exact HOME target\n'
printf '2\t{"tool_input":{"command":"git clone x ${HOME}"}}\tthe braced variable may name HOME exactly\n'
printf '2\t{"tool_input":{"command":"git clone x ~"}}\ttilde may name HOME exactly\n'
printf '2\t{"tool_input":{"command":"git worktree add $HOME topic"}}\twhitespace terminates an exact HOME target before another argument\n'
# Unquoted POSIX metacharacters terminate the target word even without spaces.
printf '2\t{"tool_input":{"command":"git clone x $HOME;echo x"}}\tsemicolon terminates an exact HOME target\n'
printf '2\t{"tool_input":{"command":"git clone x \\"$HOME\\"&& echo x"}}\tand-if terminates a quoted exact HOME target\n'
printf '2\t{"tool_input":{"command":"git clone x ${HOME}| cat"}}\ta pipe terminates a braced exact HOME target\n'
printf '2\t{"tool_input":{"command":"git clone x ~&"}}\tbackground operator terminates a tilde HOME target\n'
printf '2\t{"tool_input":{"command":"git clone x $HOME</dev/null"}}\tinput redirection terminates the target word\n'
printf '2\t{"tool_input":{"command":"git clone x $HOME>out"}}\toutput redirection terminates the target word\n'
printf '2\t{"tool_input":{"command":"( git clone x $HOME)"}}\ta subshell close terminates the exact HOME target\n'
printf '2\t{"tool_input":{"command":"git clone x $HOME\\necho x"}}\ta literal newline terminates the target word\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME_BACKUP/wt"}}\ta longer HOME-prefixed variable is a different path\n'
printf '0\t{"tool_input":{"command":"git clone x $HOMEBREW/wt"}}\tHOMEBREW is not HOME either\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME.bak/wt"}}\ta dot continues the path token into a sibling name\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME+bak/wt"}}\tplus is ordinary sibling filename content\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME@bak/wt"}}\tat-sign is ordinary sibling filename content\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME,bak/wt"}}\tcomma is ordinary sibling filename content\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME:bak/wt"}}\tcolon is ordinary sibling filename content\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME=bak/wt"}}\tequals is ordinary sibling filename content\n'
printf '0\t{"tool_input":{"command":"git clone x ${HOME}+bak/wt"}}\tbraced HOME plus suffix is still a sibling\n'
printf '0\t{"tool_input":{"command":"git clone x /home/tester+bak/wt"}}\ta literal plus-suffixed home path is a sibling\n'
printf '0\t{"tool_input":{"command":"git clone x /home/tester@bak/wt"}}\ta literal at-suffixed home path is a sibling\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME+bak/wt;echo x"}}\ta later terminator does not turn a sibling into HOME\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME@bak/wt&& echo x"}}\tand-if after a sibling preserves the allow\n'
printf '0\t{"tool_input":{"command":"git clone x \\"$HOME;bak/wt\\""}}\ta quoted semicolon is filename content, not a boundary\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME\\\\;bak/wt"}}\tan escaped semicolon is filename content, not a boundary\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME\\u001b/wt"}}\ta raw internal-marker byte is encoded as filename content\n'
printf '0\t{"tool_input":{"command":"git clone x /home/tester.bak/wt"}}\ta literal sibling path is not beneath HOME\n'
printf '0\t{"tool_input":{"command":"git clone x /home/testerx/wt"}}\ta longer literal basename is not HOME\n'
printf '0\t{"tool_input":{"command":"git clone x ~root/wt"}}\tanother account tilde is not this account HOME\n'
printf '0\t{"tool_input":{"command":"git clone x $HOME}/wt"}}\ta closing brace without an opening brace is a literal suffix\n'
printf '0\t{"tool_input":{"command":"git clone x ${HOME/wt"}}\tan opening brace without a close is not a HOME expansion\n'
# Quote removal must not create an expansion the shell never performs.
printf '0\t{"tool_input":{"command":"git clone x '"'"'$HOME'"'"'/wt"}}\tsingle-quoted HOME is a literal directory name\n'
printf '0\t{"tool_input":{"command":"git clone x \\\\$HOME/wt"}}\tan escaped dollar makes HOME literal outside quotes\n'
printf '0\t{"tool_input":{"command":"git clone x \\"\\\\$HOME\\"/wt"}}\tan escaped dollar makes HOME literal inside double quotes\n'
printf '0\t{"tool_input":{"command":"git clone x \\"~/wt\\""}}\ttilde does not expand inside double quotes\n'
printf '0\t{"tool_input":{"command":"git clone x '"'"'~/wt'"'"'"}}\ttilde does not expand inside single quotes\n'
printf '0\t{"tool_input":{"command":"git clone x \\\\~/wt"}}\tan escaped tilde is literal too\n'
printf '0\t{"tool_input":{"command":"git clone https://example.invalid/x /src/wt"}}\tcheckout onto a work filesystem is fine\n'
# Routing this arm through the shared name site also repaired an over-block it
# had carried from the start: the old whole-command regex found `git` INSIDE a
# longer word, so these two were refused at every head before this commit.
# Same class as mycurl and curl-wrapper, and refusing them is how a guard gets
# routed around instead of repaired.
printf '0\t{"tool_input":{"command":"mygit clone https://example.invalid/x $HOME/wt"}}\tmygit is a different program and its checkout is not ours\n'
printf '0\t{"tool_input":{"command":"gitfoo clone https://example.invalid/x $HOME/wt"}}\tthe name has to end where git ends\n'
printf '0\t{"tool_input":{"command":"curl -s -X GET https://git.example.invalid/api/v1/repos/a/b/pulls/1"}}\treads are never blocked\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\treview write has a wrapper\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/pulls/1/merge"}}\tmerge write has a wrapper\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d @b https://api.github.com/repos/a/b/issues"}}\tGitHub host is covered too\n'
printf '0\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/releases"}}\tan endpoint with no wrapper passes\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d {\\"event\\":\\"APPROVE\\"} https://example.invalid/x"}}\tthe APPROVE token is caught anywhere\n'
printf '0\t{"tool_input":{"command":"ls -la /src"}}\tordinary commands are untouched\n'
printf '0\t{"tool_input":{"command":"MOSAIC_WRAPPER_OVERRIDE=1 curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/pulls"}}\tbreak-glass works\n'
printf '0\t{"tool_input":{}}\tan empty payload does not block the session\n'
# --- bypasses an independent reviewer demonstrated against the first version.
# Each of these returned 0 (allowed) and each is a real write. They are pinned
# as fixtures rather than fixed-and-forgotten because the class is recurring:
# the guard reads text, so every spelling it does not know is a hole.
printf '2\t{"tool_input":{"command":"curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\t-d@body with no space is still a body\n'
printf '2\t{"tool_input":{"command":"curl --request=POST -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\t--request=POST equals-form is still a method\n'
printf '2\t{"tool_input":{"command":"p=/api/v1/repo; q=s/a/b/pulls/1/reviews; curl -d@b https://git.example.invalid${p}${q}"}}\ta path split across variables is still that path\n'
printf '2\t{"tool_input":{"command":"curl --data-binary @b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\t--data-binary is a body\n'
printf '2\t{"tool_input":{"command":"curl -F f=@b https://git.example.invalid/api/v1/repos/a/b/issues"}}\t-F multipart is a body\n'
# Reads must survive every one of those broadenings, or the guard gets disabled.
printf '0\t{"tool_input":{"command":"curl -s https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\tno body and no verb is a read\n'
printf '0\t{"tool_input":{"command":"grep -rn /pulls/ src/ | head -20"}}\ta path fragment in a grep is not an API call\n'
printf '0\t{"tool_input":{"command":"curl -X POST -d @b https://registry.example.invalid/v2/x/manifests/latest"}}\tan unwrapped API is not this guard'"'"'s business\n'
# --- round two of the same review. Splitting the ENDPOINT TOKEN defeats any
# amount of fragment matching, because the endpoint does not exist until the
# shell expands it. The guard now refuses to clear a write whose URL it cannot
# read, rather than pretending it read one.
printf '2\t{"tool_input":{"command":"a=/api/v1/repos/a/b/iss; b=ues/1/comments; curl -d@body https://git.example.invalid${a}${b}"}}\tan endpoint token split across variables is unreadable, not absent\n'
printf '2\t{"tool_input":{"command":"a=/api/v1/repos/a/b/pu; b=lls/1/reviews; curl -d@body https://git.example.invalid${a}${b}"}}\tsame split, review endpoint\n'
printf '0\t{"tool_input":{"command":"curl -X POST -d @payload https://hooks.example.invalid/services/${WEBHOOK_ID}"}}\tan opaque URL that is not forge-shaped stays allowed\n'
# --- round six changed the contract in this direction, and these fixtures are
# where it shows. They used to assert that discussing a call is not making one.
# Five rounds proved there is no textual way to tell a quoted example from a
# quoted command, so the guard stopped trying: it judges the payload, and a
# payload inside quotes is still a payload. Quoting one of these on a Bash
# command line is now refused, and the way to write the example is a
# file-writing tool. This is the deliberate cost of the mechanism change.
printf '2\t{"tool_input":{"command":"grep -R \\"curl -d https://git.example.invalid/api/v1/repos/a/b/issues\\" docs/"}}\tquoting a wrapped write is refused even in a grep\n'
printf '2\t{"tool_input":{"command":"echo \\"curl -d https://git.example.invalid/api/v1/repos/a/b/pulls\\" > note.txt"}}\t...and when written into a file\n'
printf '2\t{"tool_input":{"command":"python3 -c '"'"'print(\\"curl -d https://git.example.invalid/api/v1/repos/a/b/issues\\")'"'"'"}}\t...and when printed from another language\n'
# The boundary that keeps this from being "block everything": what is refused
# is a WRITE to a WRAPPED endpoint. Mentioning either alone still passes, and
# these are asserted as hard as the blocks above.
printf '0\t{"tool_input":{"command":"grep -R \\"curl -s https://git.example.invalid/api/v1/repos/a/b/issues/1/comments\\" docs/"}}\tquoting a READ example is untouched\n'
printf '0\t{"tool_input":{"command":"echo \\"the wrapped endpoint is https://git.example.invalid/api/v1/repos/a/b/issues/1/comments\\" >> notes.md"}}\tnaming the endpoint without a body flag is untouched\n'
printf '0\t{"tool_input":{"command":"grep -R \\"curl -d@b https://git.example.invalid/api/v1/repos/a/b/releases\\" docs/"}}\tquoting a write to an UNWRAPPED endpoint is untouched\n'
printf '0\t{"tool_input":{"command":"issue-comment.sh --repo a/b --issue 1 --body @msg.md"}}\tthe wrapper itself carries a body flag and must never trip its own guard\n'
# Command position must still catch the real thing behind operators and env.
printf '2\t{"tool_input":{"command":"cd /tmp && GITEA_TOKEN=$T curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/merge"}}\ta real call behind && and an assignment is still a call\n'
# --- the case the AUTHOR hit twice while chasing the above: sending a message
# that QUOTED one of these fixtures. Under the old contract that was a defect
# to be parsed away; under this one it is the documented cost, and the message
# gets composed with a file-writing tool instead.
printf '2\t{"tool_input":{"command":"send.sh -m \\"repro was: cd /tmp && curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/merge\\""}}\tquoting the repro in a message is refused too\n'
printf '2\t{"tool_input":{"command":"cat >> notes.md <<EOF\\nwe ran: curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues\\nEOF"}}\ta heredoc body carrying the payload is refused with it\n'
# ...but quotes stop being data the moment something executes them.
printf '2\t{"tool_input":{"command":"bash -c \\"curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/merge\\""}}\tbash -c makes the quoted text code again\n'
# --- round three. Each of these four is a real write that a bare-name match
# for the client could not see, because an ordinary word sat in front of it.
# They are kept as fixtures after the mechanism change even though the guard no
# longer looks for a client at all: they are the evidence for WHY it stopped,
# and a future re-narrowing that reintroduced position would fail here first.
printf '2\t{"tool_input":{"command":"env GITEA_TOKEN=$T curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\tenv VAR=... in front of the client is still the client\n'
printf '2\t{"tool_input":{"command":"command curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\tcommand in front of the client is still the client\n'
printf '2\t{"tool_input":{"command":"timeout 10 curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\ttimeout N in front of the client is still the client\n'
printf '2\t{"tool_input":{"command":"/usr/bin/curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\tan absolute path to the client is still the client\n'
printf '2\t{"tool_input":{"command":"echo timeout 10 curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments >> notes.md"}}\tnaming the call after echo carries the payload, so it is refused\n'
# A shell standing between quoted data and execution makes that data code,
# and the pipe is the form agents actually use. Filing it as data allowed the
# call to vanish from the skeleton while still running.
printf '2\t{"tool_input":{"command":"printf '"'"'%%s\\\\n'"'"' '"'"'curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments'"'"' | sh"}}\tquoted code piped to a shell is code\n'
printf '2\t{"tool_input":{"command":"cat <<EOF | sh\\ncurl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments\\nEOF"}}\ta heredoc piped to a shell is code\n'
printf '2\t{"tool_input":{"command":"sh -s <<EOF\\ncurl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments\\nEOF"}}\tsh -s reads its script from the heredoc\n'
# ...and the questions that used to follow — is the pipe target a shell, does a
# shell on one line execute a string on another — no longer have to be answered
# at all. Both of these carry the payload, both are refused, and neither
# outcome depends on parsing what the pipe or the other line does.
printf '2\t{"tool_input":{"command":"grep -R \\"curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments\\" docs/ | wc -l"}}\tpiping the payload to wc is refused without asking what wc is\n'
printf '2\t{"tool_input":{"command":"docker run --rm alpine sh -c '"'"'echo hi'"'"'\\necho \\"example: curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments\\" >> notes.md"}}\tan unrelated shell on another line no longer changes the answer either way\n'
# --- round four. The guard was still reading the command as typed rather than
# as the shell will run it: a backslash before a newline is removed before
# anything else happens, so the endpoint token can be split across the join.
printf '2\t{"tool_input":{"command":"curl -d@b https://git.example.invalid/api/v1/repos/a/b/iss\\\\\\nues/1/comments"}}\ta line continuation inside the endpoint token is still that endpoint\n'
printf '2\t{"tool_input":{"command":"curl -d@b https://git.example.invalid/api/v1/repos/a/b/pu\\\\\\nlls/1/reviews"}}\tsame join, review endpoint\n'
printf '2\t{"tool_input":{"command":"cat >> notes.md <<EOF\\nwe ran: curl -d@b https://git.example.invalid/api/v1/repos/a/b/iss\\\\\\nues/1/comments\\nEOF"}}\tthe join still runs first, and the joined payload is refused in a document too\n'
# Transparent prefixes take option VALUES, and the value was a word the list
# did not know — so the client went missing again behind an ordinary `sudo -u`.
printf '2\t{"tool_input":{"command":"sudo -u root curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\tan option value after a prefix does not hide the client\n'
printf '2\t{"tool_input":{"command":"timeout --signal TERM 10 curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\tan option pair plus a duration does not hide the client\n'
printf '2\t{"tool_input":{"command":"xargs echo curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\tthe payload behind xargs echo is refused rather than adjudicated\n'
printf '0\t{"tool_input":{"command":"sudo apt-get install curl"}}\tinstalling the client is not calling it\n'
# Execution through another command needed its own case under the old design.
printf '2\t{"tool_input":{"command":"find . -maxdepth 0 -exec curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments ;"}}\tfind -exec runs the client\n'
# --- round five, and the finding that ended the parser. Command substitution
# inside a double-quoted span EXECUTES, while the skeleton was discarding that
# span as inert prose. The unquoted and process-substitution forms already
# blocked, which is what made it a classification defect rather than a spelling
# one: the same call was refused or allowed depending on a quote character.
printf '2\t{"tool_input":{"command":"echo \\"$(curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments)\\""}}\tcommand substitution inside double quotes executes\n'
printf '2\t{"tool_input":{"command":"echo \\"`curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments`\\""}}\tso does the backtick form\n'
printf '2\t{"tool_input":{"command":"msg=\\"$(curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments)\\""}}\tand an assignment RHS is not data either\n'
printf '2\t{"tool_input":{"command":"echo $(curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments)"}}\tthe unquoted form, which blocked before and must keep blocking\n'
printf '2\t{"tool_input":{"command":"cat <(curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments)"}}\tprocess substitution, same\n'
printf '2\t{"tool_input":{"command":"bash --command \\"curl -d@b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments\\""}}\tthe long-option spelling of bash -c needs no entry in any list now\n'
# A client the guard was never taught is the point of dropping client
# detection: neither of these names curl at all.
printf '2\t{"tool_input":{"command":"python3 -c '"'"'import requests; requests.post(\\"https://git.example.invalid/api/v1/repos/a/b/issues/1/comments\\", json={})'"'"'"}}\ta library call is a write with no flag and no curl\n'
printf '2\t{"tool_input":{"command":"wget --post-data=x https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\twget spells its body differently and is still a write\n'
# Round six scoped the guard on `https?://`, and review found the absence shape
# had simply moved to that new boundary: a raw provider CLI carries no scheme,
# so the guard never reached the write question. These are the reported repros.
printf '2\t{"tool_input":{"command":"gh api -X POST repos/a/b/issues -f title=x -f body=y"}}\tgh api is a raw write with no URL scheme at all\n'
printf '2\t{"tool_input":{"command":"gh api -X POST repos/a/b/pulls/1/reviews -f event=APPROVE"}}\tand it reaches the endpoint the review wrapper owns\n'
printf '2\t{"tool_input":{"command":"tea api -X POST repos/a/b/issues/1/comments -f body=x"}}\ttea api, same shape, different CLI\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d x git.example.invalid/api/v1/repos/a/b/issues"}}\ta scheme-less host path is still an API write\n'
printf '2\t{"tool_input":{"command":"gh api repos/a/b/issues -f title=x"}}\tgh POSTs implicitly when handed a field, exactly as curl does with -d\n'
# ...and the boundary that stops a broader scope gate becoming block-everything.
printf '0\t{"tool_input":{"command":"gh api repos/a/b/pulls/1"}}\treading through a provider CLI stays untouched\n'
printf '0\t{"tool_input":{"command":"gh api -X POST repos/a/b/releases -f tag_name=v1"}}\tno wrapper owns releases, whoever calls it\n'
printf '0\t{"tool_input":{"command":"tea pulls create --title x --repo a/b"}}\tprovider PORCELAIN is out of scope by decision, not by accident\n'
printf '0\t{"tool_input":{"command":"rm -f /var/tmp/api/v1-issues-notes.txt"}}\t-f is only a body when it carries key=value\n'
printf '0\t{"tool_input":{"command":"grep -f patterns.txt /src/api/v1/repos/a/b/issues.log"}}\tsame, on the flag agents actually collide with\n'
# Wrong remediation is its own defect: /issues/1/labels used to block with
# "use issue-create.sh", which is not the wrapper for that call. Round seven
# answered that by letting EVERY path under a numbered issue or PR through,
# and review showed the reasoning ("no wrapper owns these") was false in this
# tree. These are the reported repros, all rc 0 before round eight, and each
# asserts the wrapper the advice must name — not merely that a block happened.
printf '2\t{"tool_input":{"command":"gh api -X PATCH repos/a/b/issues/1 -f title=x"}}\tan issue edit is issue-edit.sh, not a wrapper gap\tissue-edit.sh\n'
printf '2\t{"tool_input":{"command":"curl -X PATCH -d @b https://git.example.invalid/api/v1/repos/a/b/issues/1"}}\tsame call through curl, same wrapper\tissue-edit.sh\n'
printf '2\t{"tool_input":{"command":"gh api -X PATCH repos/a/b/issues/1/labels -f labels[]=bug"}}\tlabels are wrapped, and the advice says by which\tissue-edit.sh\n'
printf '2\t{"tool_input":{"command":"gh api -X POST repos/a/b/issues/1/assignees -f assignees[]=u"}}\tassignees are issue-assign.sh\tissue-assign.sh\n'
printf '2\t{"tool_input":{"command":"gh api repos/a/b/issues/1/assignees -f assignees[]=u"}}\tthe array field spelling is a body with no -X at all\tissue-assign.sh\n'
printf '2\t{"tool_input":{"command":"curl -X PATCH -d @b https://git.example.invalid/api/v1/repos/a/b/pulls/1/labels"}}\ta PR is an issue where labels live, so the issue wrapper owns them\tissue-edit.sh\n'
printf '2\t{"tool_input":{"command":"curl -X PATCH -d @b https://git.example.invalid/api/v1/repos/a/b/pulls/1"}}\tPR state is pr-close.sh, and the gap in that arm is stated\tpr-close.sh\n'
printf '2\t{"tool_input":{"command":"curl -X PATCH -d @b https://git.example.invalid/api/v1/repos/a/b/milestones/4"}}\ta milestone state change is milestone-close.sh, not the create wrapper\tmilestone-close.sh\n'
# SPAN. A wrapper that owns a slice of an endpoint must not be advertised as
# owning the endpoint. milestone-close.sh takes only -t <title> and sends
# state=closed, so a title/description/due-date edit is a gap and the message
# has to say so — round eight named the wrapper and stopped there.
printf '2\t{"tool_input":{"command":"curl -X PATCH -d @b https://git.example.invalid/api/v1/repos/a/b/milestones/1"}}\ta milestone edit blocks, but the advice states the close-only span\towns the CLOSE only\n'
printf '2\t{"tool_input":{"command":"gh api -X PATCH repos/a/b/issues/1 -f assignee=u"}}\tissue-edit.sh cannot set an assignee, so the message names the one that can\tissue-assign.sh owns the assignee\n'
# The residue: still genuinely owned by nothing, and still flowing through.
# Requesting a reviewer is not submitting one; pr-review.sh files verdicts and
# nothing in the tree adds a requested reviewer.
printf '0\t{"tool_input":{"command":"gh api -X POST repos/a/b/pulls/1/requested_reviewers -f reviewers[]=u"}}\tno wrapper requests a reviewer, so it is not refused with pr-review.sh\n'
# SPAN, applied to the guard's OWN fail-closed rule rather than to a wrapper.
# The scope gate admits three shapes; the unreadable-endpoint rule asked only
# for `https?://`, so a split endpoint in the other two was in scope to block,
# produced no readable endpoint, and fell through to allow. Same defect class
# as the milestone arm, one layer up. Each shape gets its own fixture, because
# a single one would have passed on the arm that already worked.
printf '2\t{"tool_input":{"command":"p=repos/a/b/iss; q=ues; gh api -X POST ${p}${q} -f title=x"}}\ta split endpoint in a provider-CLI api call is unreadable, not absent\n'
printf '2\t{"tool_input":{"command":"p=repos/a/b/issues/1/comm; q=ents; gh api -X POST ${p}${q} -f body=x"}}\tsame, comments\n'
printf '2\t{"tool_input":{"command":"p=repos/a/b/pulls/1/rev; q=iews; gh api -X POST ${p}${q} -f event=APPROVED"}}\tsame, and a verdict is the costliest one to lose\n'
printf '2\t{"tool_input":{"command":"p=repos/a/b/iss; q=ues; tea api -X POST ${p}${q} -f title=x"}}\tevery CLI the scope gate admits, not just gh\n'
printf '2\t{"tool_input":{"command":"p=/api/v1/repos/a/b/iss; q=ues; curl -X POST -d x git.example.invalid${p}${q}"}}\ta schemeless forge host with a split path is unreadable too\n'
printf '2\t{"tool_input":{"command":"h=git.example.invalid; q=ues; curl -X POST -d x ${h}/api/v1/repos/a/b/iss${q}"}}\tthe expansion may come first; the token is what matters\n'
# And the reason this is not "any variable blocks a write": a payload in a
# variable is the SAFE way to pass one and leaves the endpoint fully legible.
printf '0\t{"tool_input":{"command":"gh api repos/a/b/git/refs -f sha=$SHA"}}\tan expansion in a body value leaves the endpoint readable\n'
printf '0\t{"tool_input":{"command":"curl -X POST -d \\"$BODY\\" https://git.example.invalid/api/v1/repos/a/b/git/refs"}}\tsame for a quoted body on an unwrapped endpoint\n'
printf '0\t{"tool_input":{"command":"gh api repos/${OWNER}/${REPO}/git/refs"}}\ta read with a split endpoint is still a read\n'
printf '0\t{"tool_input":{"command":"curl -X PATCH -d @b https://git.example.invalid/api/v1/repos/a/b/issues/comments/5"}}\tediting a comment has no wrapper; only creating one does\n'
printf '0\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/issues/1/stopwatch/start"}}\tno wrapper owns a stopwatch, and none is invented for it\n'
printf '0\t{"tool_input":{"command":"gh api -X POST repos/a/b/issues/1/times -f time=60"}}\tnor time tracking\n'
printf '0\t{"tool_input":{"command":"gh api -X POST repos/a/b/issues/1/reactions -f content=+1"}}\tnor reactions\n'
# ...and the residue must be decided by the SEGMENT, never by a stray slash.
printf '2\t{"tool_input":{"command":"gh api -X PATCH repos/a/b/issues/1 -f body=see-/docs/x"}}\ta slash inside the body is not a subresource\tissue-edit.sh\n'
# The arms above the numbered ones must keep blocking, with their own wrappers.
printf '2\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/issues/1/comments"}}\tthe wrapped subresource must not fall through\tissue-comment.sh\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/issues"}}\tnor may issue creation\tissue-create.sh\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/pulls"}}\tPR creation is wrapped and must not fall through with them\tpr-create.sh\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\tand a review still names the review wrapper\tpr-review.sh\n'
# SPAN a third time, now in the scope gate itself: it asked for `/api/v[0-9]`,
# which is Gitea's spelling. GitHub's API carries no version segment at all
# (`api.github.com/repos/...`), so the schemeless Gitea write was in scope and
# the schemeless GitHub one was not — a gate calibrated to one dialect rather
# than to what identifies a provider API. `/repos/` is the marker both share.
# These endpoints are READABLE, so each asserts the wrapper it must name; a
# rc-only fixture here would pass on the unreadable arm and prove nothing.
printf '2\t{"tool_input":{"command":"curl -X POST -d x api.github.com/repos/a/b/issues"}}\ta schemeless GitHub host is a provider API even with no version segment\tissue-create.sh\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d x api.github.com/repos/a/b/issues/1/comments"}}\tsame, and the subresource still names its own wrapper\tissue-comment.sh\n'
printf '2\t{"tool_input":{"command":"host=api.github.com; curl -X POST -d x ${host}/repos/a/b/issues"}}\tthe host may be a variable; the path is what the guard reads\tissue-create.sh\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d x api.github.com/repos/a/b/pulls/1/reviews"}}\ta verdict is the costliest call to lose to a spelling\tpr-review.sh\n'
printf '2\t{"tool_input":{"command":"p=/repos/a/b/iss; q=ues; curl -X POST -d x api.github.com${p}${q}"}}\tand the split form of it is unreadable, not absent\n'
# The end-of-options marker, which is the one option not spelled like one.
printf '2\t{"tool_input":{"command":"p=repos/a/b/iss; q=ues; gh api -X POST -- ${p}${q} -f title=x"}}\ta bare -- must not walk the endpoint past the scanner\n'
# Widening a scope gate may not create a block. Reads and unwrapped endpoints
# in the newly admitted shape have to stay allowed, or this is a regression
# wearing a fix'"'"'s clothes.
printf '0\t{"tool_input":{"command":"curl api.github.com/repos/a/b/issues"}}\tadmitting a shape to the gate does not make a read a write\n'
printf '0\t{"tool_input":{"command":"curl -X POST -d x api.github.com/repos/a/b/git/refs"}}\tno wrapper owns git refs, on GitHub'"'"'s spelling either\n'
printf '0\t{"tool_input":{"command":"gh api -- repos/a/b/issues"}}\tthe marker in a read is still a read\n'
# The APPROVE trap, in the spelling a provider CLI uses, and the value that
# must never trip it.
printf '0\t{"tool_input":{"command":"curl -X POST -d {\\"event\\":\\"APPROVED\\"} https://git.example.invalid/api/v1/repos/a/b/releases"}}\tAPPROVED is the correct value and is never the trap\n'
# Documented over-block, pinned so it is a known boundary and not a surprise.
printf '2\t{"tool_input":{"command":"python3 -c '"'"'print(\\"https://git.example.invalid/api/v1/repos/a/b/issues/1/comments .post(\\")'"'"'"}}\tprose carrying .post( near a wrapped URL is refused, by the same payload rule\n'
# --- round nine, all four from one adversarial pass, and three of them are
# the same shape: a test written over the WHOLE command text deciding an
# ALLOW. That is the fail-open form this file keeps rediscovering, and it had
# reached the break-glass itself.
#
# BREAK-GLASS. `case "$CMD" in *MOSAIC_WRAPPER_OVERRIDE=1*)` cleared the entire
# command if that string appeared anywhere in it — so quoting the override in a
# note, or naming a variable after it, disabled the guard for the call sitting
# beside it. The override is now read POSITIONALLY: leading `NAME=value`
# assignments only, exactly where the shell would honour one.
printf '2\t{"tool_input":{"command":"echo \\"MOSAIC_WRAPPER_OVERRIDE=1 curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews\\" >> notes.md"}}\tquoting the override in a document does not arm it\n'
printf '2\t{"tool_input":{"command":"NOTES=MOSAIC_WRAPPER_OVERRIDE=1 curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\tan assignment whose VALUE is the override is not the override\n'
printf '2\t{"tool_input":{"command":"MOSAIC_WRAPPER_OVERRIDE=10 curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\t=10 matched the old substring test and is not the value 1\n'
printf '0\t{"tool_input":{"command":"GITEA_TOKEN=$T MOSAIC_WRAPPER_OVERRIDE=1 curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\tthe override still works behind other assignments, as the shell reads it\n'
# The cost, pinned rather than discovered later: positional means positional.
printf '2\t{"tool_input":{"command":"cd /tmp && MOSAIC_WRAPPER_OVERRIDE=1 curl -d@b https://git.example.invalid/api/v1/repos/a/b/pulls/1/merge"}}\tan override after && is not in command position and does not arm\n'
# SUBRESOURCE REFINEMENT, same defect one arm lower. It asked whether a
# subresource appears ANYWHERE in the command, so a numbered-object write was
# cleared on the strength of text in its own BODY. Inverted: clear only when
# EVERY numbered-object occurrence carries a subresource.
printf '2\t{"tool_input":{"command":"gh api -X PATCH repos/a/b/issues/1 -f body=cf-/pulls/2/files"}}\ta subresource in the body does not clear a write to the numbered issue\tissue-edit.sh\n'
printf '2\t{"tool_input":{"command":"gh api -X PATCH repos/a/b/issues/1 -f body=cf-/issues/3/reactions"}}\tsame, quoting a subresource of the same object type\tissue-edit.sh\n'
# ...and its documented cost, in the safe direction.
printf '2\t{"tool_input":{"command":"gh api -X POST repos/a/b/issues/1/reactions -f content=cf-/issues/2"}}\tan unwrapped subresource write that quotes a bare issue is refused\n'
# -K/--config. curl reads the method, the body, the headers AND the URL from
# that file, so none of them are in the command: every write test above read 0
# and the call went through. An unreadable request is not a cleared one.
printf '2\t{"tool_input":{"command":"curl --config /tmp/req https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews"}}\tthe request in a config file is unreadable, so it is refused\t--config/-K\n'
printf '2\t{"tool_input":{"command":"curl -K /tmp/req https://git.example.invalid/api/v1/repos/a/b/issues"}}\tthe short spelling, same answer\t--config/-K\n'
# Cost, stated: this refuses a --config read against a host that has nothing
# to do with a forge. The alternative is to require a forge marker in a
# command whose URL may itself be in the file, which is the hole again.
printf '2\t{"tool_input":{"command":"curl --config /tmp/req https://example.invalid/anything"}}\tan unrelated https URL with --config is refused too, by decision\t--config/-K\n'
printf '0\t{"tool_input":{"command":"eslint --config .eslintrc.json src/"}}\t--config on a command that is not curl is nobody'"'"'s business\n'
# ROUND TEN. The --config check above was first written INSIDE the API-shape
# gate, so it was guarded by a condition that the capability it guards against
# removes. A config file can carry the URL; delete the URL from the command and
# nothing is API-shaped, the branch is never entered, and the guard reports
# clean on precisely the call it exists to refuse. It is now asked of any curl.
printf '2\t{"tool_input":{"command":"curl --config /tmp/provider-write.cfg"}}\ta config file can own the URL, so there is nothing API-shaped left to gate on\t--config/-K\n'
printf '2\t{"tool_input":{"command":"curl -K/tmp/provider-write.cfg"}}\tcurl accepts the value attached to the short flag\t--config/-K\n'
printf '2\t{"tool_input":{"command":"curl -sK /tmp/provider-write.cfg"}}\tand inside a bundle, which a space-separated test does not see\t--config/-K\n'
printf '0\t{"tool_input":{"command":"tar -K /tmp/archive.tar"}}\t-K on a command that is not curl is not this hook'"'"'s business\n'
# curl by any ordinary path spelling. The first version of the config check
# matched the bare word only, so these three executed the same wrapped write
# while the guard reported clean. Recognizing only the unqualified name is
# caller-name parsing, and that is the class this file exists to refuse.
printf '2\t{"tool_input":{"command":"/usr/bin/curl --config /tmp/provider-write.cfg"}}\tan absolute path is the same invocation, not a different one\t--config/-K\n'
printf '2\t{"tool_input":{"command":"env /usr/bin/curl -K/tmp/provider-write.cfg"}}\tand it is still curl behind env, with the value attached\t--config/-K\n'
printf '2\t{"tool_input":{"command":"./curl --config /tmp/provider-write.cfg"}}\ta relative path costs two characters and used to be enough\t--config/-K\n'
# The prefix must end at a slash: a basename that merely ENDS in curl is a
# different program, and blocking it would be the over-block that gets a guard
# routed around rather than fixed.
printf '0\t{"tool_input":{"command":"mycurl --config /tmp/provider-write.cfg"}}\tmycurl is not curl, and over-blocking is its own failure\n'
printf '0\t{"tool_input":{"command":"/opt/x/curl-wrapper --config /tmp/provider-write.cfg"}}\tnor is curl-wrapper, whose name only starts the same way\n'
# And the same name once it is punctuated. The basename repair above fixed the
# UNQUOTED path spelling and nothing else, so two quote characters restored the
# bypass it had just closed: the check was still modelling one presentation of
# a shell word instead of the word. Every one of these executes the real curl.
printf '2\t{"tool_input":{"command":"\\"/usr/bin/curl\\" --config /tmp/provider-write.cfg"}}\tquoting a path does not make it a different program\t--config/-K\n'
printf '2\t{"tool_input":{"command":"'"'"'./curl'"'"' --config /tmp/provider-write.cfg"}}\tnor does quoting a relative one\t--config/-K\n'
printf '2\t{"tool_input":{"command":"$(which curl) --config /tmp/provider-write.cfg"}}\tthe name is in the text even when a substitution supplies the path\t--config/-K\n'
printf '2\t{"tool_input":{"command":"`which curl` --config /tmp/provider-write.cfg"}}\tand in the older spelling of the same substitution\t--config/-K\n'
# The provider-CLI SCOPE gate had the identical defect, untouched while the
# curl arm was repaired twice. It decides whether write detection runs at all,
# so failing to admit these is indistinguishable from allowing them — and no
# URL marker rescues them, because provider CLI paths carry no leading slash.
printf '2\t{"tool_input":{"command":"/usr/bin/gh api -X POST repos/a/b/issues -f title=x"}}\tan absolute path to a provider CLI is still a provider CLI\n'
printf '2\t{"tool_input":{"command":"./gh api -X POST repos/a/b/issues -f title=x"}}\tand a relative one still is too\n'
printf '2\t{"tool_input":{"command":"/usr/local/bin/tea api -X POST repos/a/b/issues -f title=x"}}\tthe same is true of every CLI the gate names, not just the first\n'
printf '0\t{"tool_input":{"command":"mygh api -X POST repos/a/b/issues -f title=x"}}\tmygh is not gh, and the scope gate must not over-admit either\n'
printf '0\t{"tool_input":{"command":"/usr/bin/gh api repos/a/b/issues"}}\ta read through an absolute path is still a read\n'
# Quotes and backslashes INSIDE the word. The previous repair replaced quote
# characters with whitespace, which is token separation and not quote removal:
# a shell removes a quote without splitting the word around it, so `cu"rl"` is
# one word naming curl while whitespace made it two words naming neither.
# `"/usr/bin/curl"` passed under that version only because the inserted space
# happened to land after a slash, which established nothing.
printf '2\t{"tool_input":{"command":"cu\\"rl\\" --config /tmp/provider-write.cfg"}}\ta quote inside the word does not make it another program\t--config/-K\n'
printf '2\t{"tool_input":{"command":"cu'"'"'rl'"'"' --config /tmp/provider-write.cfg"}}\tand a single quote inside it is the same word again\t--config/-K\n'
printf '2\t{"tool_input":{"command":"/usr/bin/cu\\\\rl --config /tmp/provider-write.cfg"}}\tescaping is ordinary word formation, not a disguise\t--config/-K\n'
# A backslash is NOT uniformly removed. It is literal inside single quotes,
# and inside double quotes when it precedes anything other than $, `, ",
# backslash, or newline. These spell a different program and must stay allowed.
printf '0\t{"tool_input":{"command":"'"'"'cu\\\\rl'"'"' --config /tmp/provider-write.cfg"}}\ta backslash inside single quotes remains literal\n'
printf '0\t{"tool_input":{"command":"\\"cu\\\\rl\\" --config /tmp/provider-write.cfg"}}\ta backslash before r inside double quotes remains literal\n'
printf '0\t{"tool_input":{"command":"'"'"'g\\\\it'"'"' clone https://example.invalid/x $HOME/wt"}}\ta literal backslash in a single-quoted non-git name is not a checkout\n'
printf '0\t{"tool_input":{"command":"\\"g\\\\it\\" clone https://example.invalid/x $HOME/wt"}}\ta literal backslash in a double-quoted non-git name is not a checkout\n'
# The other branch of the same rule: OUTSIDE quotes a backslash escapes the
# next character, so an escaped quote is a literal quote IN the name and the
# program is not curl. Held separately from the cases above because it is a
# different arm of the state machine, and an arm without a fixture is a rule
# that is not held.
printf '0\t{"tool_input":{"command":"cu\\\\\\"rl\\\\\\" --config /tmp/provider-write.cfg"}}\tan escaped quote is a literal quote in the name\n'
# Three quoted segments concatenate into ONE word. This is the shape that
# distinguishes quote removal from token separation, so it is worth its own line.
printf '2\t{"tool_input":{"command":"\\"cu\\"'"'"'r'"'"'\\"l\\" --config /tmp/provider-write.cfg"}}\tadjacent quoted segments are one word, and that word is curl\t--config/-K\n'
printf '2\t{"tool_input":{"command":"g\\"h\\" api -X POST repos/a/b/issues -f title=x"}}\tthe CLI name is a word on the same terms\n'
printf '2\t{"tool_input":{"command":"/usr/bin/g\\\\h api -X POST repos/a/b/issues -f title=x"}}\tincluding when it is escaped behind a path\n'
# The FLAG is the same recognition problem as the name, and is read the same
# way. No review raised this one; the name was simply the easier half to reach.
printf '2\t{"tool_input":{"command":"curl --con\\"fig\\" /tmp/provider-write.cfg"}}\tone word spelling --config is still --config\t--config/-K\n'
# The unreadable-endpoint arm is the THIRD name consumer. It kept a private
# bare-name copy of the scope gate's regex, so a caller could be admitted by
# the repaired gate and then go unrecognized by the fail-closed refinement —
# a gate and its own refinement disagreeing about who the caller is.
printf '2\t{"tool_input":{"command":"/usr/bin/gh api -X POST repos/a/b/$EP -f title=x"}}\ta path-qualified CLI with an assembled endpoint is still unreadable\n'
printf '2\t{"tool_input":{"command":"g\\"h\\" api -X POST repos/a/b/$EP -f title=x"}}\tand so is a quoted one, which is where the two halves disagreed\n'
printf '0\t{"tool_input":{"command":"mygh api -X POST repos/a/b/$EP -f title=x"}}\tmygh is still not gh, in the refinement as well as the gate\n'
# Shapes nobody raised. Written down because reasoning that they were already
# covered is precisely what produced two of the rounds above; each one below
# was measured, and the three that fail at df83a9ee are here on that evidence.
# `\curl` is the ordinary way to bypass a shell alias and is a thing people
# actually type, which makes it the least hypothetical entry in the file.
printf '2\t{"tool_input":{"command":"\\\\curl --config /tmp/provider-write.cfg"}}\tescaping the leading character to dodge an alias still names curl\t--config/-K\n'
printf '2\t{"tool_input":{"command":"cur\\"l\\" --config /tmp/provider-write.cfg"}}\tthe quote may sit at any offset in the word\t--config/-K\n'
printf '2\t{"tool_input":{"command":"g\\"h\\" api -X POST \\"repos/a/b/$EP\\" -f title=x"}}\tboth halves dressed at once, which is where they last disagreed\n'
# Over-blocking is a real failure and not a safe direction: a guard that
# refuses legitimate work gets routed around instead of repaired. These four
# pass at both heads, which is what a regression guard is for.
printf '0\t{"tool_input":{"command":"gh api \\"repos/a/b/issues\\""}}\ta quoted read is still a read\n'
printf '0\t{"tool_input":{"command":"curl https://example.com/file.txt -o /tmp/f"}}\tan ordinary download is not a provider write\n'
printf '0\t{"tool_input":{"command":"echo \\"$EP\\" && gh --version"}}\tno api subcommand, so nothing to refuse\n'
printf '0\t{"tool_input":{"command":"echo \\"not a curl call\\""}}\tthe word inside a string, with no flag, is prose\n'
# Percent-encoded endpoints. Not hypothetical: /issues/1174 and /iss%%75es/1174
# both returned HTTP 200 with the same object from the live forge, so the
# encoded spelling IS the wrapped endpoint and the literal comparison below it
# sees a segment matching nothing. Refused rather than decoded — a decoder has
# to be exactly right about depth and normalization, which is the parser
# mistake this file declines everywhere else.
printf '2\t{"tool_input":{"command":"gh api -X POST repos/a/b/iss%%75es/1/comments -f body=x"}}\tan encoded path segment reaches the wrapped endpoint\tpercent-escape\n'
printf '2\t{"tool_input":{"command":"curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/pulls/1/revi%%65ws"}}\tsame for the review endpoint, which is the one that matters most\tpercent-escape\n'
printf '2\t{"tool_input":{"command":"gh api -X POST repos/a/b/iss%%2575es/1/comments -f body=x"}}\tdouble-encoded too, which is why this refuses instead of decoding\tpercent-escape\n'
# Scoped to writes, deliberately. Reads are never blocked by this guard and a
# query string carrying %%20 is an ordinary URL, not a hazard.
printf '0\t{"tool_input":{"command":"curl -s https://git.example.invalid/api/v1/repos/a/b/issues?q=a%%20b"}}\ta percent-escape in a READ is not this hook'"'"'s business\n'
} > "$FIXTURES"
fail=0 n=0
while IFS=$'\t' read -r want payload why remedy; do
[ -n "${want:-}" ] || continue
n=$((n + 1))
out="$(printf '%s' "$payload" | "$GUARD" 2>&1)"
got=$?
if [ "$got" != "$want" ]; then
printf 'FAIL %s (want exit %s, got %s)\n' "$why" "$want" "$got"
fail=1
continue
fi
# A block that names the wrong wrapper is a defect in its own right, and until
# now it was invisible here: the harness read the exit code and nothing else,
# so /issues/1/labels blocking with "use issue-create.sh" passed every run for
# six rounds. Where a fixture states the remediation it expects, assert it.
if [ -n "${remedy:-}" ] && ! printf '%s' "$out" | grep -Fq -- "$remedy"; then
printf 'FAIL %s (blocked, but the advice does not name %s)\n' "$why" "$remedy"
fail=1
continue
fi
printf 'ok %s\n' "$why"
done < "$FIXTURES"
# ---- $HOME resolution ------------------------------------------------------
# These cannot be fixtures. Every case above varies the COMMAND; this defect
# varies the ENVIRONMENT, and the loop has no way to express that.
#
# The checkout arm built its pattern from "$HOME" without asking whether $HOME
# was a usable value. Three values it is not: unset (which is a crash under
# `set -u`, not a decision), empty (the pattern collapses to `/`, so `~|\$HOME|`
# matches whatever the empty alternative touches), and "/" (every absolute path
# is under it, so the comparison stops discriminating). An agent seat running
# with no HOME — a systemd unit without one, a container, `env -i` — got the
# checkout question answered by accident rather than on the merits.
#
# The fix resolves $HOME once, rejects all three, and BLOCKS the checkout it
# cannot adjudicate. A guard may not clear a question it was unable to ask. The
# blast radius of that fail-closed arm is asserted below to be one command shape
# and not the session: with no HOME at all, ordinary commands still pass and the
# API arms still block.
home_case() {
local why="$1" want="$2" homeval="$3" cmd="$4" needle="${5:-}"
local out got
n=$((n + 1))
local payload
payload="$(jq -nc --arg command "$cmd" '{tool_input:{command:$command}}')"
if [ "$homeval" = "@unset" ]; then
out="$(printf '%s' "$payload" | env -u HOME "$GUARD" 2>&1)"
else
out="$(printf '%s' "$payload" | env HOME="$homeval" "$GUARD" 2>&1)"
fi
got=$?
if [ "$got" != "$want" ]; then
printf 'FAIL %s (want exit %s, got %s)\n' "$why" "$want" "$got"
fail=1
return
fi
if [ -n "$needle" ] && ! printf '%s' "$out" | grep -Fq -- "$needle"; then
printf 'FAIL %s (exit %s, but the message does not say %s)\n' "$why" "$got" "$needle"
fail=1
return
fi
printf 'ok %s\n' "$why"
}
home_case 'HOME unset: a checkout is refused, not adjudicated' \
2 '@unset' 'git clone https://example.invalid/x /src/wt' 'unset or unusable'
home_case 'HOME empty: same, and it is not the same thing as unset' \
2 '' 'git clone https://example.invalid/x /src/wt' 'unset or unusable'
home_case 'HOME=/ : every path is under it, so it discriminates nothing' \
2 '/' 'git clone https://example.invalid/x /src/wt' 'unset or unusable'
home_case 'a usable HOME still allows a checkout onto a work filesystem' \
0 '/home/tester' 'git clone https://example.invalid/x /src/wt'
home_case 'a usable HOME still catches the literal path' \
2 '/home/tester' 'git clone https://example.invalid/x /home/tester/wt' 'checks a repository out under'
home_case 'a usable HOME catches the exact literal path without a trailing slash' \
2 '/home/tester' 'git clone https://example.invalid/x /home/tester' 'checks a repository out under'
home_case 'quotes around the exact literal HOME path do not change the target' \
2 '/home/tester' 'git clone https://example.invalid/x "/home/tester"' 'checks a repository out under'
home_case 'a quoted literal HOME segment remains contiguous with the suffix' \
2 '/home/tester' 'git clone https://example.invalid/x "/home/tester"/wt' 'checks a repository out under'
home_case 'and the unexpanded $HOME spelling, which needs no resolution at all' \
2 '/home/tester' 'git worktree add $HOME/wt topic' 'checks a repository out under'
# The fail-closed arm is scoped to checkouts. If it were not, a seat with no
# HOME would have every command it runs refused, which is how a guard gets
# disabled rather than fixed.
home_case 'HOME unset does not block an ordinary command' \
0 '@unset' 'ls -la /src'
home_case 'HOME unset does not stop the API arms doing their job' \
2 '@unset' 'curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews' 'pr-review.sh'
# ---- the guard standing on its own -----------------------------------------
# Every case above runs the guard from the directory holding its siblings, so
# `[ -x "$W/pr-review.sh" ]` succeeds and the $HOME fallback beside it never
# evaluates. That is a property of the HARNESS, not of the guard, and it hid a
# live fail-open: with the guard copied somewhere alone AND no HOME, the
# fallback expanded an unset variable under `set -u` and the script died at
# rc=1 — on EVERY arm, before any adjudication. A PreToolUse hook exiting
# nonzero-but-not-2 is a non-blocking error, so that seat ran with no guard and
# nothing reported it.
#
# The first remediation moved that expansion four lines earlier and called it
# closed. It was not closed, because the test could not reach it. So the guard
# is copied ALONE here — no siblings, no installed mosaic home — which is the
# deployment this file already claims to support ("still works from a repo
# checkout with no installed mosaic home").
LONE="$TMP/lone"; mkdir -p "$LONE"
cp "$GUARD" "$LONE/wrapper-guard.sh"; chmod +x "$LONE/wrapper-guard.sh"
lone_case() {
local why="$1" want="$2" homeval="$3" cmd="$4" needle="${5:-}"
local out got
n=$((n + 1))
if [ "$homeval" = "@unset" ]; then
out="$(printf '%s' "{\"tool_input\":{\"command\":\"$cmd\"}}" | env -u HOME "$LONE/wrapper-guard.sh" 2>&1)"
else
out="$(printf '%s' "{\"tool_input\":{\"command\":\"$cmd\"}}" | env HOME="$homeval" "$LONE/wrapper-guard.sh" 2>&1)"
fi
got=$?
if [ "$got" != "$want" ]; then
printf 'FAIL %s [standalone] (want exit %s, got %s)\n' "$why" "$want" "$got"
fail=1
return
fi
if [ -n "$needle" ] && ! printf '%s' "$out" | grep -Fq -- "$needle"; then
printf 'FAIL %s [standalone] (exit %s, but the message does not say %s)\n' "$why" "$got" "$needle"
fail=1
return
fi
printf 'ok %s [standalone]\n' "$why"
}
lone_case 'no siblings and no HOME: an ordinary command still passes, not rc=1' \
0 '@unset' 'ls -la /src'
lone_case 'no siblings and no HOME: a wrapped write is still refused' \
2 '@unset' 'curl -X POST -d @b https://git.example.invalid/api/v1/repos/a/b/pulls/1/reviews' 'pr-review.sh'
lone_case 'no siblings and no HOME: a checkout is refused, not adjudicated' \
2 '@unset' 'git clone https://example.invalid/x /src/wt' 'unset or unusable'
# Spelled without quotes on purpose. The payload is interpolated into a JSON
# string by the helper, so a fixture carrying bare double quotes produces
# malformed JSON, jq returns empty, and the guard exits 0 on an empty command —
# a PASS that measures nothing. That is what the first version of this case did.
lone_case 'no siblings and no HOME: the APPROVE trap still fires' \
2 '@unset' 'gh api -X POST repos/a/b/pulls/1/reviews -f event=APPROVE'
lone_case 'no siblings, usable HOME: ordinary commands unaffected' \
0 '/home/tester' 'ls -la /src'
printf '\n'
if [ "$fail" -eq 0 ]; then
printf 'wrapper-guard: %d/%d fixtures behaved as specified.\n' "$n" "$n"
else
cat <<'EOF'
wrapper-guard drifted from its contract.
A guard that blocks too much gets routed around, and a guard that blocks too
little is decoration. Both directions are failures here, which is why the
allowed cases are asserted as hard as the blocked ones.
EOF
fi
exit "$fail"