AML.T0051: LLM Prompt Injection
Override an LLM's intended behaviour via crafted prompt content.
Last reviewed October 2026
Why AML.T0051 evidence is hard
Nothing else in the ATLAS set is exercised this heavily. Overriding intended behaviour through crafted prompt content is the library's largest family: direct overrides, indirect payloads carried in documents and retrieved passages, encoding and homoglyph evasion, and the Turkish case-folding bypass that defeats a filter written in English. Findings are written with the OWASP injection identifier, and the crosswalk records the two as an exact pair.
How Penaxtra maps to AML.T0051
ATLAS is a threat-tracking taxonomy rather than a control set, so what is attached here is a technique identifier, not a pass or a fail. It lets a finding be read alongside the rest of your adversary reporting. Every finding carries the probe that was sent, a SHA-256 digest of the response it provoked, the tier that decided the verdict and its status history from first observation to close. Identity survives a re-scan: a control you closed and that comes back reopens under the same row with a reopen count, rather than arriving as a new issue.
AML.T0051 capabilities
Findings tagged with the MITRE ATLAS AML.T0051 identifier
Penaxtra severity for this control (critical)
Cross-framework identifiers attached to the same finding where controls overlap
PDF and JSON evidence export with the control identifier attached
AML.T0051 control coverage
Findings for AML.T0051 carry the MITRE ATLAS AML.T0051 identifier along with the corresponding control identifiers in the other frameworks Penaxtra maps, so one result is reflected across each mapped framework.
Frequently asked
What is AML.T0051 (LLM Prompt Injection)?
Override an LLM's intended behaviour via crafted prompt content. It is a MITRE ATLAS control; Penaxtra assesses it at critical severity.
How does Penaxtra test for AML.T0051?
ATLAS is a threat-tracking taxonomy rather than a control set, so what is attached here is a technique identifier, not a pass or a fail. It lets a finding be read alongside the rest of your adversary reporting. Every finding carries the probe that was sent, a SHA-256 digest of the response it provoked, the tier that decided the verdict and its status history from first observation to close. Identity survives a re-scan: a control you closed and that comes back reopens under the same row with a reopen count, rather than arriving as a new issue.
Does a finding for AML.T0051 help with an audit?
Each finding is tagged with the MITRE ATLAS AML.T0051 identifier and exported in the PDF and JSON evidence pack, so it appears on the auditor control list with the identifier already attached.