Follow Makefile includes when completing make targets by maxmilian · Pull Request #275 · warpdotdev/command-signatures

maxmilian · 2026-05-29T03:13:19Z

Summary

Fixes warpdotdev/warp#11705. make Tab-completion only suggested targets defined directly in the top-level Makefile; targets pulled in through include directives were invisible (other terminals show them, since GNU make treats included files as one unified ruleset).

Root cause

The make list_targets generator (command-signatures/src/generators/make.rs) ran cat [Mm]akefile, which reads only the root Makefile and never follows include / -include / sinclude directives.

Fix

Replace the command with a single self-contained awk program that echoes the root Makefile and recurses into included files, with a seen set guarding against include cycles. Design choices:

All parsing happens inside awk (not a shell loop), so behavior is identical across the user's shell — sh/bash/zsh differ in word-splitting and this sidesteps that entirely.
Injection-safe: included paths are opened directly via getline < path, never interpolated into a shell command, so a malicious Makefile can't inject commands.
Fallback: awk output is captured and only emitted on a clean exit; otherwise it falls back to the original cat [Mm]akefile. An include pointing at a real directory (an invalid Makefile that make itself rejects) aborts some awk builds with an i/o error — the fallback keeps top-level targets working instead of returning nothing.

The Rust post_process (target + ## description parsing) is unchanged.

Tests

Two end-to-end tests run the actual generator command via sh -c against temp projects:

test_list_targets_command_follows_includes — targets split across a nested include + a missing optional -include; also asserts a ## description defined in an included file survives.
test_list_targets_command_survives_directory_include — a directory include still succeeds and falls back to top-level targets.

cargo fmt --check, cargo clippy --all-targets --all-features -- -D warnings, and cargo test all pass locally (61 tests).

Scope / known limitations

Include paths that rely on globbing (include dir/*.mk) or make variables (include $(VAR)) are intentionally left unresolved — expanding them safely would require either a shell (injection risk) or evaluating the Makefile. Explicit file paths (the reported case and the common case) and arbitrarily nested includes are handled.

lucieleblanc · 2026-06-05T16:37:48Z

/oz-review

oz-for-oss · 2026-06-05T16:38:07Z

@lucieleblanc

I'm re-reviewing this pull request in response to a review request.

You can view the conversation on Warp.

I completed the review and no human review was requested for this pull request.

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss

Overview

This PR changes the make command completion generator to recursively print Makefile include contents before the existing target parser runs.

Concerns

The new include follower opens paths supplied by the repository's Makefile without bounding them to the project, so completion can read and parse local files outside the repo.

Security

A malicious repository can add an absolute or parent-traversal include path and cause tab completion to read and parse local files before the user runs make. Restrict recursive include following to project-local files or skip paths that resolve outside the working tree.

Verdict

Found: 0 critical, 1 important, 0 suggestions

Request changes

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss · 2026-06-05T16:44:45Z

+/// Known limitation: include paths that rely on globbing (`include dir/*.mk`) or make variables
+/// (`include $(VAR)`) are not expanded, since resolving them safely would require either a shell
+/// (injection risk) or evaluating the Makefile.
+const LIST_TARGETS_COMMAND: &str = r##"out=$(awk 'function process(f,  line,rest,n,i,parts){if(f in seen)return;seen[f]=1;while((getline line<f)>0){print line;if(match(line,/^[ \t]*[-s]?include[ \t]+/)){rest=substr(line,RLENGTH+1);n=split(rest,parts,/[ \t]+/);for(i=1;i<=n;i++)if(parts[i]!="")process(parts[i])}}close(f)}BEGIN{for(i=1;i<ARGC;i++)process(ARGV[i]);exit}' [Mm]akefile 2>/dev/null) && printf '%s\n' "$out" || cat [Mm]akefile 2>/dev/null"##;


⚠️ [IMPORTANT] [SECURITY] This follows Makefile-controlled include paths without bounding them to the project, so a malicious repo can make tab completion read and parse arbitrary local files (for example absolute paths or .. traversals). Resolve include paths and skip anything outside the working tree before getline.

maxmilian · 2026-06-06T01:07:36Z

Thanks @lucieleblanc / Oz — addressed the security concern (the one important finding).

Fix: the recursive include-follower now confines itself to the project subtree. A new safe() guard in the awk program skips any include path that is absolute (/etc/…), home-relative (~/…), or contains a .. component, so a hostile repository’s Makefile can no longer make tab-completion read and parse files outside the working tree (e.g. include /etc/passwd or include ../../.ssh/id_rsa) before the user ever runs make.

This is a safe boundary because GNU make resolves include paths relative to its working directory (not the including file), so a legitimate project-local include is always a relative, non-.. path — those keep working unchanged. Absolute/../~ includes are intentionally not followed (documented as a known limitation; their targets just aren’t surfaced in completion).

Added a regression test (test_list_targets_command_ignores_includes_outside_project) that places real, readable sentinel .mk files outside the project dir, references them via both an absolute path and a ../ path, and asserts neither target surfaces while the in-tree target still does — proving the generator chooses not to follow them.

Local cargo fmt --check / cargo clippy --all-targets -D warnings / full test suite (72 passing, 4 in the make module) all green; rebased onto latest main.

/oz-review

oz-for-oss · 2026-06-06T01:07:53Z

@maxmilian

I'm re-reviewing this pull request in response to a review request.

You can view the conversation on Warp.

I completed the review and no human review was requested for this pull request.

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss

Overview

This PR replaces make target completion input with an awk-based include expander and adds end-to-end tests for included makefiles and defensive fallback cases.

Concerns

The include expander does not fully enforce the stated project-subtree boundary: the path check is lexical, and tab-indented recipe lines can still be mistaken for include directives.

Security

A hostile repository can use a relative symlink or an include-looking recipe line to make tab completion read files outside the intended Makefile include graph before the user runs make.

Verdict

Found: 0 critical, 1 important, 0 suggestions

Request changes

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss · 2026-06-06T01:11:31Z

+/// (injection risk) or evaluating the Makefile. Absolute, home-relative, and `..`-escaping
+/// includes are intentionally not followed (see the security boundary above); their targets are
+/// simply not surfaced in completion.
+const LIST_TARGETS_COMMAND: &str = r##"out=$(awk 'function safe(p){return p!~/^\//&&p!~/^~/&&p!~/(^|\/)\.\.(\/|$)/}function process(f,  line,rest,n,i,parts){if(f in seen)return;seen[f]=1;while((getline line<f)>0){print line;if(match(line,/^[ \t]*[-s]?include[ \t]+/)){rest=substr(line,RLENGTH+1);n=split(rest,parts,/[ \t]+/);for(i=1;i<=n;i++)if(parts[i]!=""&&safe(parts[i]))process(parts[i])}}close(f)}BEGIN{for(i=1;i<ARGC;i++)process(ARGV[i]);exit}' [Mm]akefile 2>/dev/null) && printf '%s\n' "$out" || cat [Mm]akefile 2>/dev/null"##;


⚠️ [IMPORTANT] [SECURITY] This scanner still follows paths Make would not include: safe() is only lexical, so a project-local symlink can point outside the tree, and the regex accepts tab-prefixed recipe lines as include directives. Reject symlinks/out-of-tree realpaths and avoid treating recipe lines as directives before calling getline.

maxmilian · 2026-06-06T01:32:02Z

Addressed both points from the latest review.

1. Recipe-line over-match. The include-directive match was ^[ \t]* (allowed a leading tab). A tab-indented line in make is a recipe (a shell command), not a directive, so a \tinclude foo recipe line could be followed. Anchored to spaces only (^ *…); a genuine space-indented include still works.

2. Symlink escape (lexical check insufficient). Correct — safe() is purely lexical, so an in-tree symlink (evil -> /etc) with include evil/passwd is textually in-tree but resolves outside. I reworked the command from the all-in-awk recursion into a POSIX-sh visit() walker that resolves each candidate with realpath and follows it only when it stays under $(pwd -P):

visit(){ rp=$(realpath -- "$1" 2>/dev/null)||return 0;
  case "$rp" in "$root"|"$root"/*) ;; *) return 0;; esac;   # symlink-resolved containment
  …cat -- "$rp"; recurse into safe() include paths… }

awk now only extracts include paths (never opens them); every path reaches the shell as a quoted argument, never interpolated, so there is still no command-injection surface. If realpath is unavailable the command falls back to cat [Mm]akefile (top-level only) rather than following includes unsafely. Containment is now two layers: lexical safe() (abs / ~ / ..) + realpath symlink resolution.

Tests (6 in the make module, all green): nested includes + missing optional include (descriptions survive), directory include, absolute + .. escapes against real out-of-tree sentinels, a tab-indented recipe include line, and an in-tree symlink resolving out of tree — each asserts only in-tree targets surface. cargo fmt --check / clippy --all-targets -D warnings / full suite all green on latest main.

/oz-review

oz-for-oss · 2026-06-06T01:32:20Z

@maxmilian

I'm re-reviewing this pull request in response to a review request.

You can view the conversation on Warp.

I reviewed this pull request and requested human review from: @lucieleblanc.

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

Oz no longer requests changes for this pull request after the latest automated review.

oz-for-oss

Overview

This updates the make target generator to expand project-local included Makefiles while preserving a containment boundary for untrusted repositories, and adds end-to-end tests covering nested includes, invalid directory includes, path traversal, recipe-indented directives, and symlink escapes.

Concerns

Include operand parsing currently does not strip trailing Make comments before word-splitting, so comment words can be followed as extra include paths if matching files exist.

Verdict

Found: 0 critical, 0 important, 1 suggestions

Approve with nits

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss · 2026-06-06T01:36:35Z

+/// safely would require a shell (injection risk) or evaluating the Makefile. Absolute,
+/// home-relative, `..`-escaping, and symlink-escaping includes are intentionally not followed (see
+/// the security boundary above); their targets are simply not surfaced in completion.
+const LIST_TARGETS_COMMAND: &str = r##"root=$(pwd -P)||exit 0;command -v realpath >/dev/null 2>&1||{ cat [Mm]akefile 2>/dev/null;exit 0;};seen="|";visit(){ rp=$(realpath -- "$1" 2>/dev/null)||return 0;case "$rp" in "$root"|"$root"/*) ;; *) return 0;; esac;case "$seen" in *"|$rp|"*) return 0;; esac;seen="$seen$rp|";cat -- "$rp" 2>/dev/null;set -f;for inc in $(awk 'function safe(p){return p!~/^\//&&p!~/^~/&&p!~/(^|\/)\.\.(\/|$)/} /^ *[-s]?include[ \t]+/{match($0,/^ *[-s]?include[ \t]+/);rest=substr($0,RLENGTH+1);n=split(rest,parts,/[ \t]+/);for(i=1;i<=n;i++)if(parts[i]!=""&&safe(parts[i]))print parts[i]}' "$rp" 2>/dev/null);do set +f;visit "$inc";set -f;done;set +f;};for f in [Mm]akefile;do [ -e "$f" ]&&visit "$f";done"##;


💡 [SUGGESTION] Strip trailing Make comments before splitting include operands; otherwise include common.mk # comment treats # and comment words as additional include paths when matching files exist, which can surface targets make would not load.

The `make` `list_targets` generator ran `cat [Mm]akefile`, which reads only the top-level Makefile, so targets defined in files pulled in via `include`/`-include`/`sinclude` were never suggested (warpdotdev/warp#11705). Replace the command with a POSIX-sh `visit()` walker over the include graph: it `cat`s a file, uses a small awk program to extract that file's include paths, and recurses into each. awk only extracts paths (it never opens them) and every path is passed to the shell as a quoted argument, never interpolated, so a Makefile cannot inject commands. The extractor only matches `include` directives indented with spaces, not a leading tab -- a tab-indented line is a make recipe (a shell command), so a tab-indented `include`-looking line is deliberately not followed -- and strips a trailing `# comment` before splitting so comment words are not mistaken for include paths. Confine include-following to the project subtree so a hostile repository cannot make tab-completion read files outside the working tree (e.g. `include /etc/passwd`) before the user runs `make`, using two layers: - a lexical `safe()` guard that skips absolute, home-relative, and `..`-escaping include paths; and - `realpath`, which canonicalizes each remaining candidate so it is followed only when it resolves under `$(pwd -P)`. This stops escapes a lexical check cannot, e.g. an in-tree symlink (`evil -> /etc`) used as `include evil/passwd`. GNU make resolves include paths relative to its working directory, so legitimate project-local includes are unaffected. A `seen` set of canonical paths guards against include cycles. If `realpath` is unavailable the command falls back to `cat [Mm]akefile` (top-level only) rather than following includes unsafely. Glob (`include dir/*.mk`), variable (`include $(VAR)`), and whitespace-containing include paths are intentionally left unresolved. Add end-to-end tests run against temp projects: nested includes plus a missing optional include (asserting `##` descriptions survive and a trailing-comment operand is not followed), a directory include, an absolute and a `..`-escaping include pointing at real out-of-tree sentinels, a tab-indented (recipe) `include`-looking line, and an in-tree symlink resolving out of tree -- asserting only in-tree targets surface in each.

maxmilian · 2026-06-06T01:39:10Z

Addressed the trailing-comment nit. The include extractor now strips a trailing # comment from the directive line before word-splitting, so a comment word is never followed as an include path (added a regression case to test_list_targets_command_follows_includes: include makefiles/main.mk # makefiles/leak.mk with a real makefiles/leak.mk present, asserting its target does not surface).

6 make-module tests + cargo fmt --check / clippy --all-targets -D warnings all green on latest main.

/oz-review

oz-for-oss · 2026-06-06T01:39:28Z

@maxmilian

I'm re-reviewing this pull request in response to a review request.

You can view the conversation on Warp.

I reviewed this pull request and requested human review from: @lucieleblanc.

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss

Overview

This PR updates the make target generator to follow include, -include, and sinclude directives with a shell/awk walker, while constraining traversal to files that canonicalize under the current directory. It also adds end-to-end tests for nested includes, optional missing includes, directory includes, path escape attempts, recipe-indented include-like lines, and symlink escapes.

Concerns

Non-blocking: the command currently concatenates each visited makefile with no guaranteed separator, so a file without a final newline can merge with the next included file's first line and corrupt target parsing.

Verdict

Found: 0 critical, 0 important, 1 suggestions

Approve with nits

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss · 2026-06-06T01:53:29Z

+/// safely would require a shell (injection risk) or evaluating the Makefile. Absolute,
+/// home-relative, `..`-escaping, and symlink-escaping includes are intentionally not followed (see
+/// the security boundary above); their targets are simply not surfaced in completion.
+const LIST_TARGETS_COMMAND: &str = r##"root=$(pwd -P)||exit 0;command -v realpath >/dev/null 2>&1||{ cat [Mm]akefile 2>/dev/null;exit 0;};seen="|";visit(){ rp=$(realpath -- "$1" 2>/dev/null)||return 0;case "$rp" in "$root"|"$root"/*) ;; *) return 0;; esac;case "$seen" in *"|$rp|"*) return 0;; esac;seen="$seen$rp|";cat -- "$rp" 2>/dev/null;set -f;for inc in $(awk 'function safe(p){return p!~/^\//&&p!~/^~/&&p!~/(^|\/)\.\.(\/|$)/} /^ *[-s]?include[ \t]+/{match($0,/^ *[-s]?include[ \t]+/);rest=substr($0,RLENGTH+1);sub(/#.*/,"",rest);n=split(rest,parts,/[ \t]+/);for(i=1;i<=n;i++)if(parts[i]!=""&&safe(parts[i]))print parts[i]}' "$rp" 2>/dev/null);do set +f;visit "$inc";set -f;done;set +f;};for f in [Mm]akefile;do [ -e "$f" ]&&visit "$f";done"##;


💡 [SUGGESTION] Add an explicit newline separator after each cat; if a makefile lacks a trailing newline, the next included file starts on the same output line and list_targets_post_process can miss or corrupt that first target.

`visit()` cat'd each makefile in the include graph back-to-back with no separator, so a file without a trailing newline merged its last line into the first line of the next included file -- a target at the top of an included file was swallowed into the preceding recipe line and dropped from completion. Emit an explicit newline (`echo`) after each `cat`. The extra blank line between files is inert to the target parser. Add a regression test: an include reached from a newline-less root Makefile whose last line is a recipe must still surface its leading target.

maxmilian · 2026-06-19T09:03:19Z

Thanks @oz-for-oss — addressed the last suggestion (the missing separator between concatenated makefiles).

Fix: visit() now emits an explicit newline (echo) after each cat, so a makefile without a trailing newline can no longer merge its last line into the next included file's first line. The extra blank line between files is inert to the target parser.

Regression test: test_list_targets_command_separates_files_without_trailing_newline builds a newline-less root Makefile whose last line is a recipe and includes a file whose leading target would otherwise be swallowed into that recipe line; it asserts the included target still surfaces — red before this change, green after. Full suite: 65 passed, cargo fmt --check clean, cargo clippy --all-targets clean.

/oz-review

oz-for-oss · 2026-06-19T09:03:38Z

@maxmilian

I ran into an unexpected error while working on this.

Powered by Oz

maxmilian · 2026-06-19T09:07:42Z

@lucieleblanc — this one's been waiting on human review for a while; I think it's ready whenever you have a moment.

Where it landed:

Oz's last review was Approve with nits (1 non-blocking suggestion).
I've now addressed that final suggestion too — the missing newline separator between concatenated makefiles. Without it, a makefile lacking a trailing newline could merge its last line into the next included file's first line, dropping that file's leading target from completion. Fix + a red→green regression test pushed in f4ceae0.
Full suite green: 65 passed, cargo fmt --check and cargo clippy --all-targets both clean.

(I re-triggered Oz to re-confirm, but it hit an internal error on its end — happy to re-run it if you'd like a fresh automated pass before merging.)

Thanks!

maxmilian · 2026-06-27T06:15:07Z

Gentle follow-up @lucieleblanc 🙂 — no rush, just flagging in case this slipped off the radar. #275 has been green and ready for a few weeks now: Oz approved-with-nits, that final nit addressed, full suite passing (cargo fmt/clippy clean). Happy to re-trigger a fresh Oz pass first if that'd help before merge.

maxmilian · 2026-08-05T08:37:03Z

Re-triggering Oz, because the last attempt never actually produced a review.

On Jun 19 I pushed f4ceae0 (the newline separator between concatenated makefiles — Oz's final non-blocking suggestion) and re-triggered. Oz answered "I ran into an unexpected error while working on this" and no review followed. So the newest review on this PR is still the Jun 6 Approve with nits, which predates the commit that fixes the nit it raised. Anyone reading the review state sees an outstanding suggestion that has in fact been addressed for seven weeks.

Nothing about the change has moved since, and the gap it fixes is still open on main — I re-checked today: command-signatures/src/generators/make.rs (165 lines) has no include handling of any kind, and #281 is still open.

Where the four Oz findings landed, for whoever picks this up:

Finding	Resolution
[IMPORTANT/SECURITY] unbounded include paths	`safe()` guard rejecting absolute, `~`-relative, and `..` paths
[IMPORTANT/SECURITY] lexical check misses symlinks; tab-indented recipe lines matched as directives	rewritten as a POSIX-sh `visit()` walker resolving each candidate with `realpath`; directive match anchored to spaces only
[SUGGESTION] trailing `# comment` treated as include operands	comment stripped before word-splitting, with a regression case
[SUGGESTION] no separator between concatenated makefiles	explicit newline after each `cat`, red→green regression test (`f4ceae0`)

Full suite green: 65 passed, cargo fmt --check and cargo clippy --all-targets -D warnings clean.

/oz-review

oz-for-oss · 2026-08-05T08:37:23Z

@maxmilian

I'm re-reviewing this pull request in response to a review request.

You can view the conversation on Warp.

I completed the review and no human review was requested for this pull request.

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

oz-for-oss

Overview

This PR updates the make target completion generator to follow explicit Makefile include directives while bounding traversal to files that canonicalize under the working tree. It also adds end-to-end coverage for nested includes, optional/missing includes, directory includes, path escape attempts, recipe-indented include-looking lines, symlink escapes, and newline-less file concatenation.

Concerns

No blocking concerns found in the annotated diff.

Verdict

Found: 0 critical, 0 important, 0 suggestions

Approve

Comment /oz-review on this pull request to retrigger a review (up to 3 times on the same pull request).

Powered by Oz

maxmilian · 2026-08-05T09:01:11Z

Oz came back clean on the re-trigger: 0 critical, 0 important, 0 suggestions — Approve. So the review state now matches the code, which it hadn't since June.

One thing I noticed comparing that run to the earlier ones, and it may explain why this PR has been quiet rather than anyone ignoring it: this run ended with "no human review was requested for this pull request", whereas the Jun 6 runs ended with "requested human review from @lucieleblanc". If routing is what's missing, that's not something I can do from here — it would just need someone to pick it up or point it at the right reviewer.

No action needed from me that I can see, but happy to rebase, re-run anything, or split it up if that would help.

cla-bot Bot added the cla-signed label May 29, 2026

maxmilian mentioned this pull request May 29, 2026

Autocomplete for make doesn't work when Makefile uses include warpdotdev/warp#11705

Open

4 tasks

maxmilian force-pushed the maxmilian/fix-11705-make-includes branch 2 times, most recently from adb42f0 to 1c60a56 Compare May 29, 2026 03:22

maxmilian marked this pull request as ready for review May 29, 2026 03:24

lucieleblanc requested review from lucieleblanc and oz-agent June 3, 2026 17:19

This comment was marked as duplicate.

Sign in to view

lucieleblanc linked an issue Jun 5, 2026 that may be closed by this pull request

Autocomplete for make doesn't work when Makefile uses include directive #281

Open

oz-for-oss Bot mentioned this pull request Jun 5, 2026

Autocomplete for make doesn't work when Makefile uses include directive #281

Open

oz-for-oss Bot previously requested changes Jun 5, 2026

View reviewed changes

warpdotdev deleted a comment from oz-for-oss Bot Jun 5, 2026

maxmilian force-pushed the maxmilian/fix-11705-make-includes branch from 1c60a56 to 7028708 Compare June 6, 2026 01:07

oz-for-oss Bot previously requested changes Jun 6, 2026

View reviewed changes

maxmilian mentioned this pull request Jun 6, 2026

Normalize Windows drive-letter SSH Include paths in host completions #280

Closed

maxmilian force-pushed the maxmilian/fix-11705-make-includes branch from 7028708 to dd79529 Compare June 6, 2026 01:31

oz-for-oss Bot reviewed Jun 6, 2026

View reviewed changes

maxmilian force-pushed the maxmilian/fix-11705-make-includes branch from dd79529 to 0de986a Compare June 6, 2026 01:39

oz-for-oss Bot reviewed Jun 6, 2026

View reviewed changes

oz-for-oss Bot reviewed Aug 5, 2026

View reviewed changes

Uh oh!

Conversation

maxmilian commented May 29, 2026

Summary

Root cause

Fix

Tests

Scope / known limitations

Uh oh!

This comment was marked as duplicate.

Uh oh!

This comment was marked as duplicate.

Uh oh!

lucieleblanc commented Jun 5, 2026

Uh oh!

oz-for-oss Bot commented Jun 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oz-for-oss Bot left a comment

Choose a reason for hiding this comment

Overview

Concerns

Security

Verdict

Uh oh!

oz-for-oss Bot Jun 5, 2026

Choose a reason for hiding this comment

Uh oh!

maxmilian commented Jun 6, 2026

Uh oh!

oz-for-oss Bot commented Jun 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oz-for-oss Bot left a comment

Choose a reason for hiding this comment

Overview

Concerns

Security

Verdict

Uh oh!

oz-for-oss Bot Jun 6, 2026

Choose a reason for hiding this comment

Uh oh!

maxmilian commented Jun 6, 2026

Uh oh!

oz-for-oss Bot commented Jun 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oz-for-oss Bot left a comment

Choose a reason for hiding this comment

Overview

Concerns

Verdict

Uh oh!

oz-for-oss Bot Jun 6, 2026

Choose a reason for hiding this comment

Uh oh!

maxmilian commented Jun 6, 2026

Uh oh!

oz-for-oss Bot commented Jun 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

oz-for-oss Bot left a comment

Choose a reason for hiding this comment

Overview

Concerns

Verdict

Uh oh!

oz-for-oss Bot Jun 6, 2026

Choose a reason for hiding this comment

Uh oh!

maxmilian commented Jun 19, 2026

Uh oh!

oz-for-oss Bot commented Jun 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

maxmilian commented Jun 19, 2026

Uh oh!

maxmilian commented Jun 27, 2026

Uh oh!

oz-for-oss Bot commented Jun 5, 2026 •

edited

Loading

oz-for-oss Bot commented Jun 6, 2026 •

edited

Loading

oz-for-oss Bot commented Jun 6, 2026 •

edited

Loading

oz-for-oss Bot commented Jun 6, 2026 •

edited

Loading

oz-for-oss Bot commented Jun 19, 2026 •

edited

Loading

oz-for-oss Bot commented Aug 5, 2026 •

edited

Loading