Sample mastery report
Illustrativeevidence: a seeded worked example, not your organisation’s data and not a measured or certified result.
Globex: AI Security Engineer mastery
Where this seat sits on the skills it owns, which gaps still hold the climb, and the curated next skills that close them. Read in a few minutes; act on one gap.
- Prepared on
- The aisecurityskills personal mastery check
- Basis
- Cloud Security Alliance (CSA) AI Controls Matrix (AICM) controls · Mythos-style four rungs
- Evidence state
- Illustrative · fictional organisation
- Issued
- 18 Jul 2026
Industrial conglomerate · fictional organisation · illustrative sample
The verdict
Specialist split: model security leads; agent containment and attack-path drills still open.
Globex's AI Security Engineer seat reads 5 implemented · 17 partial · 4 absent across 26 prompts across 5 signature groups (secure / attack / contain / runtime / data). 21 skills still sit below target, led by agent containment. The specialist path is proving integrity through the lifecycle — and bounding what agents may do.
Implemented
5
26 prompts · 5 groups
Partial / Absent
17 · 4
Below optimum on the personal bank
Below target
21
Personal gate open on at least one autonomy-holding control
The ladder
Four concrete rungs from initial to optimum. Implemented, Partial, and Absent tallies are derived from the same session answers the live mastery check uses.
- 5ImplementedAt the personal optimum for that control
- 17PartialIn progress - not yet claimable as done
- 4AbsentNot practiced or not in place for this seat
Your personal gate is open: at least one autonomy-holding control sits below its target. Close those first; the same controls that gate the org climb.
Next skills
Prescriptions draw only from the curated core skill set via prescribeMasteryGaps. Each closes a derived gap on this seat's session bank.
- 01Execution
Implementing Llm Guardrails For Security
Closes the AIS-11 gap (Draw and enforce security boundaries around agents): Absent → Partial.
implementing-llm-guardrails-for-security
- 02Execution
Gating Irreversible Agent Actions With Human Approval
Closes the IAM-17 gap (Gate output-changing powers behind special authorization): Absent → Partial.
gating-irreversible-agent-actions-with-human-approval
- 03Execution
Bounding Agent Autonomy And Tool Scopes Least Privilege
Closes the IAM-18 gap (Restrict what agents may access, by design): Absent → Partial.
bounding-agent-autonomy-and-tool-scopes-least-privilege
- 04Execution
Performing Web Application Penetration Test
Closes the TVM-07 gap (Penetration-test systems that include AI): Absent → Partial.
performing-web-application-penetration-test
- 05Judgment
Authoring Model Cards And System Documentation
Closes the MDS-05 gap (Validate model documentation against reality): Partial → Implemented.
authoring-model-cards-and-system-documentation