← Incident Database
Indirect Prompt InjectionHigh
Slack AI private-channel data exfiltration
August 2024 · Slack AI
What happened
An attacker who could post in any public channel could plant instructions that Slack AI later executed for a victim with private-channel access, rendering a Markdown link that leaked private content such as an API key to an attacker server.
Root cause
Slack AI ingested public-channel content as trusted context and rendered attacker-controlled Markdown links, enabling exfiltration without the attacker accessing the private data directly.
Fix / outcome
Slack investigated and deployed a patch, stating it found no evidence of unauthorized customer-data access.
Sources
Learn this attack class
This incident is an example of Indirect Prompt Injection. Read the guide, then try it hands-on in the Academy.