Codestin Search App

johnstcn · 2026-07-06T17:30:57Z

Adds a search parameter to GetChats. A chat matches when the search hits its title (FTS), its PR title (FTS), a message body (FTS over the backfilled chat_messages.search_tsv), or its exact PR number when the search is all digits. Empty search is a no-op and composes with the existing archived, repo, and pr_status filters. A ^[0-9]{1,18}$ guard keeps the PR-number cast from erroring or overflowing on oversized digit strings.

Stacked on #26995, which stacks on #26968; those add the schema migration and the BackfillChatMessagesSearchTsv query.

Refs https://linear.app/codercom/issue/CODAGT-724

EXPLAIN smoke check (20k seeded chats, messages, and diff statuses). The partial GIN indexes serve the message and PR-title arms:

SubPlan 2
  ->  Bitmap Heap Scan on chat_diff_statuses cds_1 (actual time=1.913..1.913 rows=0 loops=1)
        Recheck Cond: (to_tsvector('simple'::regconfig, pull_request_title) @@ '''kubernetes'''::tsquery)
        ->  Bitmap Index Scan on idx_chat_diff_statuses_pr_title_fts (actual time=1.911..1.911 rows=0 loops=1)
SubPlan 4
  ->  Bitmap Heap Scan on chat_messages cm_1 (actual time=1.696..1.696 rows=0 loops=1)
        Recheck Cond: ((search_tsv @@ '''kubernetes'''::tsquery) AND (NOT deleted) AND ...)
        ->  Bitmap Index Scan on idx_chat_messages_search_tsv (actual time=1.695..1.696 rows=0 loops=1)

Cross-table OR arms force a sequential scan of chats for the title arm, but idx_chats_title_fts serves the title predicate itself:

Bitmap Heap Scan on chats (actual time=0.669..0.858 rows=20 loops=1)
  ->  Bitmap Index Scan on idx_chats_title_fts (actual time=0.660..0.660 rows=40 loops=1)
        Index Cond: (to_tsvector('simple'::regconfig, title) @@ '''kubernetes'''::tsquery)

🤖 This PR was generated by Coder Agents on behalf of @johnstcn.

linear-code · 2026-07-06T17:31:02Z

CODAGT-724

johnstcn · 2026-07-08T13:03:38Z

/coder-agents-review

coder-agents-review · 2026-07-08T13:03:42Z

Chat: Review posted | View chat
Requested: 2026-07-09 07:23 UTC by @johnstcn
Spend: $74.89 / $100.00

Review history

R1 (2026-07-08): 16 reviewers, 2 Nit, 3 Note, 1 P3, COMMENT. Review
R2 (2026-07-08): 10 reviewers, 5 Nit, 3 Note, 2 P3, COMMENT. Review
R3 (2026-07-08): 10 reviewers, 9 Nit, 3 Note, 3 P3, COMMENT. Review
R4 (2026-07-09): 10 reviewers, 9 Nit, 3 Note, 3 P3, APPROVE. Review

deep-review v0.9.0 | Round 4 | 01a47d9..7b927dd

Last posted: Round 4, 15 findings (3 P3, 9 Nit, 3 Note), APPROVE. Review

Finding inventory

Findings

#	Sev	Status	Location	Summary	Round	Reviewer	Posted
CRF-1	P3	Author fixed (`98a120b`)	chats.sql:674	Comment claims AND-of-terms but websearch_to_tsquery also supports OR/NOT; misattributes case-insensitivity to the function rather than 'simple' config	R1	Ryosuke P3, Leorio Note, Gon P2	Yes
CRF-2	Nit	Author fixed (`98a120b`)	chats.sql:679	SQL arm labels restate what code shows; index references are the real value	R1	Gon P2	Yes
CRF-3	Nit	Author fixed (`98a120b`)	querier_test.go:15435	Test comment first clause restates test name; trim to the trap	R1	Gon P2	Yes
CRF-4	Note	Author fixed (`98a120b`)	querier_test.go:15417	Test section headers (// 1. ... // 11.) repeat what test names convey	R1	Gon	Yes
CRF-5	Note	Author fixed (`98a120b`)	chats.sql:705	AND short-circuit for bigint cast relies on PostgreSQL behavior, not SQL standard	R1	Hisoka	Yes
CRF-6	Note	Author accepted R2 (no code change; 'simple' semantics intentional, docs deferred to CODAGT-727)	chats.sql:680	'simple' config means no stemming; "auth" won't match "authentication"	R1	Ryosuke	Yes
CRF-7	Note	Dropped by orchestrator (stacked PR with linked ticket CODAGT-724)	queries.sql.go:8776	Search field no production caller	R1	Kite P4	No
CRF-8	Nit	Dropped by orchestrator (factually incorrect; no blank line exists at line 693)	chats.sql:693	Blank line in EXISTS subquery	R1	Kite	No
CRF-9	P3	Author fixed (`f14b054`)	chats.sql:677	Whitespace-only search returns zero rows; empty search returns all rows	R2	Hisoka	Yes
CRF-10	Nit	Author fixed (`f14b054`)	chats.sql:673	SQL comments contain sentences that restate what code shows (header at 673, message-body at 689, PR-number at 702)	R2	Gon P2	Yes
CRF-11	Nit	Author fixed (`f14b054`)	querier_test.go:15272	Test separators (--- helpers --- at 15272, --- fixtures --- at 15356) carry no category	R2	Gon P2	Yes
CRF-12	Nit	Author fixed (`f14b054`)	querier_test.go:15382	Test comments contain leading clauses that narrate setup (backfill at 15382, delete at 15387)	R2	Gon P2	Yes
CRF-13	P3	Author fixed (`7b927dd`)	querier_test.go:15365	Message search positive path covers only (user, both); assistant role and user visibility not tested on root chats	R3	Bisky	Yes
CRF-14	Nit	Author fixed (`7b927dd`)	chats.sql:675	btrim comment restates guard and is inaccurate (btrim strips spaces, not all whitespace)	R3	Gon P2, Leorio Note	Yes
CRF-15	Nit	Author fixed (`7b927dd`)	querier_test.go:15391	Comment leading clause restates the SQL UPDATE	R3	Gon P2	Yes
CRF-16	Nit	Author fixed (`7b927dd`)	querier_test.go:15376	modelMsg named by visibility while toolMsg named by role	R3	Gon	Yes
CRF-17	Nit	Author fixed (`7b927dd`)	querier_test.go:15425	IneligibleRolesNoMatch name says Roles but covers both role and visibility	R3	Gon	Yes

Round log

Round 1

Panel (16 reviewers). 1 P3, 2 Nit new. 3 Notes. 2 dropped. Reviewed against f367661..d06ddf4.

Round 2

Panel (10 reviewers). CRF-1 through CRF-5 addressed, CRF-6 accepted. 1 P3, 3 Nit new. Reviewed against 554ef39..2a4e560.

Round 3

Panel (10 reviewers). CRF-9 through CRF-12 addressed. 1 P3, 4 Nit new. Reviewed against 01a47d9..4904714.

Round 4

Panel (10 reviewers). CRF-13 through CRF-17 addressed. No new findings. Convergence. Reviewed against 01a47d9..7b927dd.

About deep-review

CRF = Coder Review Finding (P0-P4, Nit, Note)

Reviewer	Focus
Bisky	tests
Chopper	ops/errors
Churn-guard	change verification
Ging	language modernization
Gon	naming
Hisoka	edge cases
Killua	perf
Kite	change integrity
Knov	contracts
Knuckle	SQL
Komugi	flake/determinism
Kurapika	security
Law	decomposition
Leorio	docs
Luffy	product
Mafu-san	process
Mafuuu	contracts
Melody	dispatch/pairing
Meruem	structural
Nami	frontend
Netero	mechanical checks
Pariston	premise testing
Pen-botter	product gaps
Razor	verification
Robin	duplication
Ryosuke	Go arch
Takumi	concurrency
Zoro	shape

🤖 Managed by Coder Agents.

johnstcn · 2026-07-08T21:42:33Z

/coder-agents-review

coder-agents-review

All R1 findings addressed. CRF-1 comment rewritten with accurate websearch_to_tsquery docs. CRF-5's AND short-circuit replaced with CASE/WHEN per the SQL standard. CRF-2/3/4 comment cleanup applied. CRF-6 accepted with docs deferred to CODAGT-727.

1 P3, 3 Nit. No blockers.

"The forced search_tsv update on toolMsg and modelMsg is the real test. Without the manual UPDATE the test would pass by accident." (Bisky)

Process note: the fix commit subject ("refactor(coderd/database): address review on GetChats search") names the process rather than the change. Something like "guard bigint cast with CASE in GetChats search" would tell a future git blame reader what changed.

🤖 This review was automatically generated with Coder Agents.

coder-agents-review · 2026-07-08T22:12:11Z

+    -- PR number. websearch_to_tsquery accepts quoted phrases, OR, and
+    -- -negation; the 'simple' config folds case and skips stemming.
+    AND CASE
+        WHEN @search::text != '' THEN (


P3 [CRF-9] Whitespace-only search returns zero rows; empty search returns all rows.

"" hits ELSE true and returns everything. " " passes != '', enters the THEN branch, and websearch_to_tsquery('simple', ' ') produces an empty tsquery that matches nothing. All four OR arms fail, zero rows returned.

Input Rows

"" all

" " 0

"-" 0

"OR" 0

"'' vs ' ' is a sharp discontinuity at the SQL layer. Every future caller inherits it." (Hisoka)

Fix: BTRIM(@search) != '' or LENGTH(BTRIM(@search)) > 0. The - and OR cases are arguably correct (operators without operands), but whitespace producing zero rows when empty produces all rows is a trap.

(Hisoka P3)

🤖

issue(fixed): Guard is now btrim(@search::text) != '', so whitespace-only search returns all rows, same as empty. Pinned by WhitespaceSearch/ReturnsAll alongside the existing EmptySearch cases. Punctuation-only inputs ("OR", "-") still return 0 rows here by FTS semantics; the API layer rejects them with 400 via PR #27093's preflight. Fixed in f14b054.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

coder-agents-review · 2026-07-08T22:12:11Z

        )
        ELSE true
    END
+    -- Free-text search across chat title, PR title, message bodies, and


Nit [CRF-10] The first sentence ("Free-text search across chat title, PR title, message bodies, and PR number") enumerates what the four OR arms below already show. Same pattern at line 689 ("Rows pending backfill (search_tsv IS NULL) match nothing" restates the IS NOT NULL predicate) and line 702 ("The length cap prevents overflow" restates the regex). Each comment has a useful trap sentence; the trailing restatement could be trimmed.

(Gon)

🤖

chore: Trimmed all three restatements; each comment now carries only its trap. Fixed in f14b054.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

coder-agents-review · 2026-07-08T22:12:11Z

+	})
+	require.NoError(t, err)
+
+	// --- helpers ---


Nit [CRF-11] The --- helpers --- and --- fixtures --- (line 15356) separators are navigational aids in a long test function, but carry no invariant, trap, or contract. The helper names (createRoot, createChild, insertMsg, linkPR) and the subsequent variable declarations are self-documenting.

(Gon)

🤖

chore: Separators deleted. Fixed in f14b054.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

coder-agents-review · 2026-07-08T22:12:11Z

+	toolMsg := insertMsg(ineligibleChat.ID, database.ChatMessageRoleTool, database.ChatMessageVisibilityBoth, "forbidden secret token")
+	modelMsg := insertMsg(ineligibleChat.ID, database.ChatMessageRoleUser, database.ChatMessageVisibilityModel, "forbidden secret token")
+
+	// Backfill search_tsv through the real pipeline. The backfill indexes


Nit [CRF-12] The leading clauses at lines 15382 ("Backfill search_tsv through the real pipeline") and 15387 ("Deleting after backfill removes the row from search results") narrate the test setup sequence. The trap sentences that follow are the real value. Trimming to:

// Ineligible rows keep search_tsv NULL after backfill.

// Soft-deleted rows stay excluded even though search_tsv remains populated.

keeps the invariant without the sequence narration.

(Gon)

🤖

chore: Took both one-liners nearly verbatim. Fixed in f14b054.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

johnstcn · 2026-07-08T23:02:44Z

/coder-agents-review

All prior findings addressed and replied to inline. Requesting a convergence pass.

🤖 Requested via Coder Agents on behalf of @johnstcn.

coder-agents-review

All R2 findings addressed. CRF-9 whitespace guard landed with btrim. CRF-10/11/12 comment cleanup applied.

1 P3, 4 Nit. No blockers.

"If someone narrowed the SQL to cm.role = 'user' or cm.visibility = 'both', every test in this suite would still pass." (Bisky)

🤖 This review was automatically generated with Coder Agents.

coder-agents-review · 2026-07-08T23:26:55Z

+	linkPR(mergedChat.ID, "https://github.com/acme/other-repo/pull/7", "merged", "Fix authentication flow", 7, "https://github.com/acme/other-repo.git")
+
+	msgChat := createRoot("plain one")
+	insertMsg(msgChat.ID, database.ChatMessageRoleUser, database.ChatMessageVisibilityBoth, "kubernetes cluster restart")


P3 [CRF-13] The message search positive path covers only one of four eligible (role, visibility) pairs.

The SQL accepts cm.role IN ('user', 'assistant') AND cm.visibility IN ('user', 'both'), giving four eligible combinations. Every positive message match on a root chat uses (user, both):

msgChat at line 15365: ChatMessageRoleUser, ChatMessageVisibilityBoth

The assistant role only appears on childChat (line 15372), where the assertion tests child exclusion, not role eligibility. The user visibility (distinct from both) also only appears there.

"If someone narrowed the SQL to cm.role = 'user' or cm.visibility = 'both', every test in this suite would still pass. The exclusion tests prove that tool role and model visibility are rejected, but they don't prove that assistant role or user visibility are accepted." (Bisky)

Fix: add one message with (assistant, both) and one with (user, user) on root chats, each with a distinct search term, and add table cases asserting they match.

(Bisky P3)

🤖

issue(fixed): Added root-chat fixtures for the three missing eligible pairs: (assistant, both), (user, user), and (assistant, user), each with a distinct term and a positive table case. Narrowing role or visibility in the SQL now fails the suite. Fixed in 7b927dd.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

coder-agents-review · 2026-07-08T23:26:55Z

    END
+    -- websearch_to_tsquery accepts quoted phrases, OR, and -negation;
+    -- the 'simple' config folds case and skips stemming. btrim makes
+    -- whitespace-only search behave like empty.


Nit [CRF-14] "btrim makes whitespace-only search behave like empty" restates the btrim(@search::text) != '' guard on the next line. Additionally, PostgreSQL btrim(text) strips ASCII space (U+0020) only, not tabs or newlines, so "whitespace-only" overstates the scope (practical impact is zero, but the comment is imprecise).

Dropping the sentence resolves both issues:

-- websearch_to_tsquery accepts quoted phrases, OR, and -negation; -- the 'simple' config folds case and skips stemming.

(Gon P2, Leorio Note)

🤖

issue(fixed): Dropped the sentence, and fixed the gap it papered over: the guard is now btrim(@search::text, E' \t\n\r') != '', so tab/newline-only input behaves like empty too. Pinned by TabOnlySearch/ReturnsAll. Fixed in 7b927dd.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

coder-agents-review · 2026-07-08T23:26:55Z

+	pendingChat := createRoot("plain four")
+	insertMsg(pendingChat.ID, database.ChatMessageRoleUser, database.ChatMessageVisibilityBoth, "elasticsearch indexing")
+
+	// Force search_tsv onto ineligible rows to prove the role and


Nit [CRF-15] "Force search_tsv onto ineligible rows" restates the SQL UPDATE that follows. The why-not-what is the second half. Trimming to:

// Prove role/visibility predicates exclude rows even when search_tsv is set.

(Gon)

🤖

chore: Took your one-liner. Fixed in 7b927dd.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

coder-agents-review · 2026-07-08T23:26:55Z

+
+	ineligibleChat := createRoot("plain three")
+	toolMsg := insertMsg(ineligibleChat.ID, database.ChatMessageRoleTool, database.ChatMessageVisibilityBoth, "forbidden secret token")
+	modelMsg := insertMsg(ineligibleChat.ID, database.ChatMessageRoleUser, database.ChatMessageVisibilityModel, "forbidden secret token")


Nit [CRF-16] toolMsg is named by role (tool) but modelMsg is named by visibility (model-only). A reader scanning the pair assumes symmetric naming. modelOnlyMsg or modelVisMsg would surface the visibility dimension.

(Gon)

🤖

chore: Renamed to modelOnlyMsg. Fixed in 7b927dd.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

coder-agents-review · 2026-07-08T23:26:55Z

+		{"Message/DeletedNoMatch", database.GetChatsParams{Search: "terraform"}, nil},
+		// Parent also excluded: EXISTS is per-chat, not per-tree.
+		{"Message/ChildNotSurfaced", database.GetChatsParams{Search: "orchestrator saga"}, nil},
+		{"Message/IneligibleRolesNoMatch", database.GetChatsParams{Search: "forbidden secret"}, nil},


Nit [CRF-17] IneligibleRolesNoMatch says "Roles" but the test covers both role exclusion (toolMsg) and visibility exclusion (modelMsg). IneligibleMessagesNoMatch would be more accurate.

(Gon)

🤖

chore: Renamed to IneligibleMessagesNoMatch. Fixed in 7b927dd.

🤖 Reply posted by Coder Agents on behalf of @johnstcn.

johnstcn · 2026-07-09T07:23:09Z

/coder-agents-review

R3 findings addressed in 7b927dd, including a behavioral upgrade of CRF-14 (btrim now strips tabs/newlines, pinned by test).

🤖 Requested via Coder Agents on behalf of @johnstcn.

coder-agents-review

All 17 findings resolved across 4 rounds. The PR is clean.

R1 caught comment inaccuracies (websearch_to_tsquery semantics, case-insensitivity attribution) and a SQL-standard compliance gap (AND short-circuit for bigint cast). R2 found the whitespace/empty search discontinuity. R3 closed the (role, visibility) test coverage gap and tightened comments. R4 confirmed convergence: 10 reviewers, zero new findings.

The author went beyond fixes on two occasions: CRF-5 was treated as correctness rather than style (CASE for guaranteed evaluation order), and CRF-14's comment drop led to widening btrim to strip tabs/newlines with a pinning test.

Process note (repeated from R2): the commit eb282b97b9 subject "address review" names the process, not the change. Future git blame readers benefit from subjects that describe what changed.

"These comments explain what would break if you changed the code without understanding it. That is what comments are for." (Leorio)

🤖 This review was automatically generated with Coder Agents.

github-actions Bot assigned johnstcn Jul 6, 2026

johnstcn force-pushed the cian/codagt-722-dbpurge-sweep-to-backfill-and-maintain branch from e11a3a8 to f367661 Compare July 8, 2026 10:57

johnstcn force-pushed the cian/codagt-724-add-search-parameter-and-recency-ordering-to-getchats branch from de5d61f to d06ddf4 Compare July 8, 2026 10:58

johnstcn mentioned this pull request Jul 8, 2026

DNM: feat(coderd/searchquery): parse search filter for chats #27090

Closed

johnstcn mentioned this pull request Jul 8, 2026

DNM: feat: wire chat search filter into chats API #27093

Closed

This comment was marked as resolved.

Sign in to view

johnstcn force-pushed the cian/codagt-724-add-search-parameter-and-recency-ordering-to-getchats branch from d06ddf4 to 98a120b Compare July 8, 2026 14:28

johnstcn force-pushed the cian/codagt-722-dbpurge-sweep-to-backfill-and-maintain branch from 7381fb1 to 2d5bcc9 Compare July 8, 2026 14:51

johnstcn force-pushed the cian/codagt-724-add-search-parameter-and-recency-ordering-to-getchats branch from 98a120b to 71399d8 Compare July 8, 2026 14:52

johnstcn force-pushed the cian/codagt-722-dbpurge-sweep-to-backfill-and-maintain branch from 2d5bcc9 to c6a884b Compare July 8, 2026 18:38

johnstcn force-pushed the cian/codagt-724-add-search-parameter-and-recency-ordering-to-getchats branch from 71399d8 to f17ee99 Compare July 8, 2026 18:39

johnstcn force-pushed the cian/codagt-722-dbpurge-sweep-to-backfill-and-maintain branch from c6a884b to 7ee0e38 Compare July 8, 2026 19:16

johnstcn force-pushed the cian/codagt-724-add-search-parameter-and-recency-ordering-to-getchats branch from f17ee99 to 30d3b97 Compare July 8, 2026 19:17

johnstcn force-pushed the cian/codagt-722-dbpurge-sweep-to-backfill-and-maintain branch from 7ee0e38 to 554ef39 Compare July 8, 2026 19:33

johnstcn force-pushed the cian/codagt-724-add-search-parameter-and-recency-ordering-to-getchats branch from 30d3b97 to 2a4e560 Compare July 8, 2026 19:34

coder-agents-review Bot reviewed Jul 8, 2026

View reviewed changes

johnstcn added 4 commits July 8, 2026 22:42

feat(coderd/database): add search parameter to GetChats

676cb16

docs(coderd/database): tighten search comments

3e7d1d2

refactor(coderd/database): address review on GetChats search

eb282b9

fix(coderd/database): treat whitespace-only GetChats search as empty

4904714

johnstcn force-pushed the cian/codagt-724-add-search-parameter-and-recency-ordering-to-getchats branch from f14b054 to 4904714 Compare July 8, 2026 22:43

coder-agents-review Bot reviewed Jul 8, 2026

View reviewed changes

test(coderd/database): cover all eligible GetChats search message pairs

7b927dd

coder-agents-review Bot approved these changes Jul 9, 2026

View reviewed changes

johnstcn changed the title ~~feat(coderd/database): add search parameter to GetChats~~ DNM: feat(coderd/database): add search parameter to GetChats Jul 9, 2026

johnstcn mentioned this pull request Jul 9, 2026

feat: Add full text search over chat messages #27126

Merged

github-actions Bot added the stale This issue is like stale bread. label Jul 19, 2026

github-actions Bot closed this Jul 23, 2026

Uh oh!

Conversation

johnstcn commented Jul 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

linear-code Bot commented Jul 6, 2026

Uh oh!

johnstcn commented Jul 8, 2026

Uh oh!

coder-agents-review Bot commented Jul 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Findings

Round log

Round 1

Round 2

Round 3

Round 4

Uh oh!

This comment was marked as resolved.

Uh oh!

johnstcn commented Jul 8, 2026

Uh oh!

coder-agents-review Bot left a comment

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

johnstcn commented Jul 8, 2026

Uh oh!

coder-agents-review Bot left a comment

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

coder-agents-review Bot Jul 8, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

johnstcn commented Jul 9, 2026

Uh oh!

coder-agents-review Bot left a comment

Choose a reason for hiding this comment

Uh oh!

johnstcn commented Jul 6, 2026 •

edited

Loading

coder-agents-review Bot commented Jul 8, 2026 •

edited

Loading