SQLite FTS5 Index Corruption from Dirty Data
SkillDatabases & dataFix SQLite "database disk image is malformed" errors caused by corrupted FTS5 indexes. Use when: (1) UPDATE/INSERT fails with "stepping, database disk image is malformed" but PRAGMA integrity_check returns "ok", (2) Only specific records fail to update while others succeed, (3) Records contain dirty text with tabs, newlines, or unusual characters that were inserted into FTS5-indexed columns. The misleading error message suggests disk corruption, but the actual issue is FTS5 index corruption from malformed text content.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the SQLite FTS5 Index Corruption from Dirty Data skill
What this skill tells your AI
The instructions your AI receives, as published by divinevideo/divine-mobile in .agents/skills/sqlite-fts5-corruption-rebuild/SKILL.md and read by ahel’s review.
Problem
SQLite returns "database disk image is malformed" (error code 11) when trying to
UPDATE or INSERT records, even though PRAGMA integrity_check reports the database
as "ok". The error only affects specific records, not all writes.
Context / Trigger Conditions
- Error message:
Error: stepping, database disk image is malformed (11) PRAGMA integrity_checkreturnsok(misleading!)- SELECT queries on the same records work fine
- Only certain records fail to update; other records in the same table update successfully
- The affected records contain text with embedded tabs (
\t), newlines (\n), or other control characters in columns that are indexed by FTS5 - A concurrent process (like a Node.js server) may have the database open
Solution
-
Identify FTS5 tables linked to the affected table:
SELECT name FROM sqlite_master WHERE type='table' AND name LIKE '%fts%'; -
Stop any processes holding the database open (the rebuild requires exclusive access):
kill $(lsof -ti :PORT) # Kill the server process -
Rebuild the FTS5 index:
INSERT INTO table_fts(table_fts) VALUES('rebuild');Replace
table_ftswith your actual FTS5 table name (e.g.,people_fts). -
Retry the failed operation - it should now succeed.
-
Restart your application server.
Verification
- The previously failing UPDATE/INSERT statement now succeeds
- Full-text search queries still return correct results
- No more "malformed" errors on subsequent writes
Example
# Error occurs:
sqlite3 data/tracker.db "UPDATE people SET name = 'More' WHERE id = 11;"
# Error: stepping, database disk image is malformed (11)
# But integrity check passes:
sqlite3 data/tracker.db "PRAGMA integrity_check;"
# ok
# And SELECT works fine:
sqlite3 data/tracker.db "SELECT id, name FROM people WHERE id = 11;"
# 11|More Masajes\t\n\t\tEscort 23 años...
# Fix: Stop server, rebuild FTS5 index
kill $(lsof -ti :3001)
sqlite3 data/tracker.db "INSERT INTO people_fts(people_fts) VALUES('rebuild');"
# Now the update works:
sqlite3 data/tracker.db "UPDATE people SET name = 'More' WHERE id = 11;"
# Success!
Prevention
- Sanitize text before inserting into FTS5-indexed columns:
const cleanName = rawName.replace(/[\t\n\r]+/g, ' ').trim(); - Strip control characters from scraped data before database insertion
- Consider using
content=(external content) FTS5 tables that can be rebuilt independently without affecting the main table
Notes
- The error code 11 (
SQLITE_CORRUPT) is the same as actual disk corruption, making this hard to diagnose - The corruption is in the FTS5 shadow tables (
_data,_idx,_docsize), not the main table, which is whyintegrity_checkpasses - If
rebuildfails with "database is locked", ensure no other process has the database open (check withlsoforfuser) - This can also happen when records are deleted or updated externally (e.g., via sqlite3 CLI) while a WAL-mode connection is active
References
Signals
- GitHub stars
- 265
- Forks
- 55
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
sqlite-fts5-corruption-rebuild- Source
- github.com/divinevideo/divine-mobile