Forum

Forums
OpenClaw
Credential and Secr...
Credential Leakage ...
 
Notifications
Clear all

Credential Leakage via Agents and Logs

Cases where OpenClaw agents expose credentials through tool call outputs, LLM responses, or log files — analysis, mitigations, and detection patterns.
RSS
Page 1 / 4

How to: Make your own tools that return opaque handles instead of raw data.

Last post by Morgan T., 1 month ago

Posts: 3
Views: 0


Beginner question: Should I run my agent in a sandbox just for this?

Last post by Kira Freak, 1 month ago

Posts: 4
Views: 0


Anyone else having agent outputs full of AWS session tokens in plain text?

Last post by Sofia Lindgren, 1 month ago

Posts: 1
Views: 0


Practical guide: Setting up eBPF to monitor for credential-like data exfiltration

Last post by Erin V., 1 month ago

Posts: 4
Views: 1


Step-by-step: Isolating agent network traffic to catch unexpected calls.

Last post by Chloe Nakamura, 1 month ago

Posts: 1
Views: 0


Breaking: New post on the OpenClaw blog about 'defense in depth' for agents

Last post by Oliver James, 1 month ago

Posts: 4
Views: 1


Thoughts on the new 'sanitize output' flag in Claw 2.1?

Last post by Bill Cartwright, 1 month ago

Posts: 1
Views: 1


Why does my PDF parsing tool output the embedded metadata with API keys?

Last post by Alice Wye, 1 month ago

Posts: 4
Views: 1


Hot take: Default configurations exist to make demos easy, not to be secure.

Last post by Mary K., 1 month ago

Posts: 1
Views: 1


Help: My agent is outputting database connection strings when it errors.

Last post by Emma R., 1 month ago

Posts: 1
Views: 2


Page 1 / 4
Share:
Share
Tweet
Share