BoundaryVault guides: agent log forensics, in plain language
Three short guides on reading an AI agent's action log after something went wrong — how credential grabs are found after the fact, what scope escape actually means, and how a disclosure packet is built. Every guide ends at the same place: the free log scan, which runs the real detector on your own log.
Agent Log Forensics: Finding Credential Grabs After the Fact
A credential grab in an agent log leaves traces: secret-shaped values, key-named arguments, decoded payloads. How BoundaryVault's detector finds them.
What Is Scope Escape in an AI Agent?
Scope escape is an agent acting outside its declared allowed-tool set. How BoundaryVault defines it, detects it, and why undeclared scope cannot be judged.
Building a Disclosure Packet After an Agent Incident
After an agent incident, BoundaryVault builds a review-only disclosure packet: summary, patterns, timeline, redacted evidence, rotation worklist. The steps.
Start with your own log
The guides describe the detector; the scan runs it. Paste a recorded agent action log and get the findings — credential grabs, token forges, scope escapes, wallet touches — with redacted evidence, free.
More from the fleet: Ingenious AI builds the full set of agent-security tools.