Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Deploys and monitors ransomware canary files across critical directories using Python's watchdog library for real-time filesystem event detection. Places strategically named decoy files that mimic high-value targets (financial records, credentials, database exports) in locations ransomware typically enumerates first. Monitors for any read, modify, rename, or delete operations on canary files and triggers immediate alerts via email, Slack webhook, or syslog when interaction is detected, providing early warning before full encryption begins.
.claude/skills/deploying-ransomware-canary-files/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-12 | ✗→✓ | ▲ Improved | — | — |
| case-04 | ✗→✓ | ▲ Improved | — | — |
| case-01 | ✗→✓ | ▲ Improved | — | — |
| case-03 | ✗→✓ | ▲ Improved | — | — |
| case-02 | ✗→✓ | ▲ Improved | — | — |
Do not use as a replacement for endpoint protection, backup strategy, or network segmentation. Canary files are a detection layer, not a prevention mechanism.
Create decoy files with realistic names and content that attract ransomware scanners. Files should have names like Passwords.xlsx, Financial_Report_2026.docx, backup_credentials.csv and contain plausible-looking but fake data. Place them in directories ransomware typically targets first: user desktops, Documents folders, network share roots, and backup paths.
Use Python's watchdog library with a custom FileSystemEventHandler that watches canary file paths. The handler triggers on on_modified, on_deleted, on_moved, and on_created events for canary files. Any legitimate user or process should never touch these files, so any interaction is a high-confidence indicator of ransomware or unauthorized access.
Wire the filesystem monitor to multiple alert channels: email via SMTP, Slack webhook POST, syslog forwarding to SIEM, and local log file. Include the triggering event type, file path, timestamp, and process information (when available) in alert payloads.
Simulate ransomware behavior by programmatically modifying, renaming, and deleting canary files to verify the detection pipeline fires correctly. Measure time-to-alert and validate alert delivery across all configured channels.
| Term | Definition | |------|------------| | Canary File | A decoy file placed in a monitored directory that triggers an alert when accessed, modified, or deleted | | Watchdog | Python library that monitors filesystem events using OS-native APIs (inotify on Linux, FSEvents on macOS, ReadDirectoryChangesW on Windows) | | Honey File | Synonym for canary file; a fake document designed to attract and detect malicious activity | | Entropy Check | Measuring randomness in file content to detect encryption (ransomware produces high-entropy output) |
RANSOMWARE CANARY ALERT
========================
Timestamp: 2026-03-11T14:23:07Z
Event: FILE_MODIFIED
Canary File: /srv/shares/finance/Passwords.xlsx
Directory: /srv/shares/finance
SHA-256 Before: a3f2...8b4c
SHA-256 After: 7e91...2d3f
Alert Channels: [email, slack, syslog]
Action: Investigate immediately - potential ransomware activity| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-13 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-12 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-06 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-22 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-04 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-01 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-14 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-10 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-03 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-08 | pass→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-07 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-18 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-17 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-11 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-02 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-09 | pass→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-16 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-15 | fail→pass | — | — | — | — | — | — | — | — | — | — | — | — |
case-21 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-05 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-19 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
case-20 | fail→fail | — | — | — | — | — | — | — | — | — | — | — | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +32 percentage points is the difference between those two pass rates over the 22 comparable cases.
The per-case answers from this run were removed by the retention sweep, so the case table below shows the verdicts without the text either arm produced. The counts above were recorded at the time and are unaffected. Answers are now kept for 180 days.
Other measured skills in the registry, with their headline benchmark lift.