Head to head
Aider vs Cursor
Comparing 5 documented Aider incidents against 4 for Cursor.
What these records show
Aider has the lower average severity among the incidents collected here (4.0 vs 7.8 for Aider and Cursor, from 5 and 4 incidents). These records can't establish which agent is safer: counts reflect how widely each is used and how often people report failures, not a failure rate.
| Metric | Aider | Cursor |
|---|---|---|
| Documented incidents | 5 | 4 |
| Average severity | 4.0 | 7.8 |
| Critical | 0 | 1 |
| High | 0 | 3 |
| Source checked | 2 | 4 |
Severity at a glance
Failure modes
The incidents behind these numbers
Aider
6.0Aider leaks model-provider API keys to child processes spawned via /run, /test, lint commands, and /git4.0Aider's headless `--message` mode silently swallows slash commands (`/model`, `/ask`, `/architect`) and exits 0 having done nothing4.0Aider exits with code 0 after a fatal API connection failure, masking hard failures as successful headless runs3.0Aider's unified-diff coder silently drops the partial-application warning when one hunk succeeds and another fails3.0Aider silently discards a model's correct diff response as "no tracked changes" when it guesses the wrong edit format
Cursor
9.0Cursor's terminal sandbox trusted an agent-set working directory, letting zero-click prompt injection escape it and gain code execution (CVE-2026-50548)8.0Malicious cloned repository triggered code execution in Cursor on Windows7.0Cursor agent approved to delete one 18.5GB temp folder instead ran a mis-quoted rmdir that wiped a developer's source, Git repo, and six months of ERP backups7.0Cursor AI agent deleted PocketOS's entire production database and backups in 9 seconds