Asi sqlite-scanner
Scan filesystems for SQLite databases by magic-byte detection. Wraps simonw/sqlite-scanner (Go binary via PyPI/uvx). Use for forensic inventory, honeypot canary detection, VM disk auditing, and database cataloging.
git clone https://github.com/plurigrid/asi
T=$(mktemp -d) && git clone --depth=1 https://github.com/plurigrid/asi "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/sqlite-scanner" ~/.claude/skills/plurigrid-asi-sqlite-scanner && rm -rf "$T"
skills/sqlite-scanner/SKILL.mdsqlite-scanner - SQLite Database Filesystem Scanner
Overview
sqlite-scanner wraps Simon Willison's Go CLI tool that identifies SQLite databases by reading the first 16 bytes of every file and comparing against the magic header
SQLite format 3\x00. No file extension guessing — pure binary signature detection.
Role: MINUS verifier in triadic consensus — validates filesystem state by detecting hidden/renamed SQLite databases.
Quick Start
# One-shot via uvx (no install needed) uvx sqlite-scanner ~/ # With JSON output and sizes uvx sqlite-scanner --json --size /path/to/scan # Streaming JSONL for pipeline consumption uvx sqlite-scanner --jsonl --size ~/Library ~/Documents
When to Use
- Forensic inventory: Find all SQLite databases on a system regardless of extension
- VM disk auditing: Scan mounted VM disk images for database artifacts
- Honeypot validation: Verify canary SQLite files are planted correctly
- Pre-migration cataloging: Inventory databases before system moves
- Stealer artifact detection: Find browser/app databases in unexpected locations
When NOT to Use
- Querying database contents (use DuckDB or sqlite3)
- Modifying databases (use appropriate write tools)
- Scanning network shares (latency makes magic-byte reads slow)
CLI Reference
| Flag | Default | Description |
|---|---|---|
| | Parallel worker goroutines |
| | Include file size in output |
| | Pretty-printed JSON array |
| | Newline-delimited JSON |
| - | Print version |
Positional args: One or more directories to scan. Defaults to
. if none given.
Output Formats
Plain text (default)
/Users/bob/Library/Safari/History.db /Users/bob/Library/Messages/chat.db
JSON (--json --size
)
--json --size{"entries": [ {"path": "/Users/bob/Library/Safari/History.db", "size": 1048576}, {"path": "/Users/bob/Library/Messages/chat.db", "size": 524288} ]}
JSONL (--jsonl --size
)
--jsonl --size{"path":"/Users/bob/Library/Safari/History.db","size":1048576} {"path":"/Users/bob/Library/Messages/chat.db","size":524288}
Detection Method
var sqliteMagic = []byte("SQLite format 3\x00") func checkSQLiteMagic(path string) bool { f, _ := os.Open(path) buf := make([]byte, 16) io.ReadFull(f, buf) return bytes.Equal(buf, sqliteMagic) }
Worker-pool pattern:
filepath.WalkDir feeds paths channel, N workers read 16 bytes each, matches stream to output immediately.
GF(3) Conservation
sqlite-scanner is assigned trit = -1 (MINUS) for verification role:
Verifier (-1) + Coordinator (0) + Generator (+1) = 0 (mod 3) [sqlite-scanner] [jo-clojure] [hy-regime]
Triad semantics:
- sqlite-scanner detects/validates SQLite presence (MINUS)
- jo-clojure orchestrates scan → query → report pipeline (ERGODIC)
- hy-regime generates analysis reports from discovered databases (PLUS)
Boxxy Activities
scan-host
Scan host filesystem for all SQLite databases.
(sqlite-scanner/scan ["/Users/bob"] {:json true :size true :workers 8})
scan-vm-disk
Mount and scan a VM disk image.
(sqlite-scanner/scan-mounted disk-mount-path {:jsonl true})
canary-audit
Verify honeypot canary databases are in place.
(sqlite-scanner/verify-canaries canary-paths scan-results)
Architecture
- upstream:
(single main.go, ~250 LOC)github.com/simonw/sqlite-scanner - distribution: PyPI wheels via
(8 platform targets)go-to-wheel - invocation:
(zero-install) oruvx sqlite-scannergo install - concurrency: Go worker pool,
default workersruntime.NumCPU() - integration: Joker
activities wrap via.jokejoker.os/exec
References
- Blog post — Distribution pattern
- GitHub — Source
- PyPI — Package
- go-to-wheel — Build tool