Scope and publication standards
Training disclosures
Read what the training includes, how its sources are checked and what its previews and credentials mean.
Practical training
sicurity.ai provides practitioner-authored training for the AI Security Engineer. The core pathway secures AI systems; specialization and adjacent courses have separate scopes. Course pages describe planned or proposed modules, practical work and release status.
Study-time basis
2 to 4 hours per module, plus practice, is a planning estimate, not a measured completion time. Additional practice time varies. Training does not guarantee a pass, employment or employer recognition.
Previews, data and exercise scope
Interface images are design previews with sample data, not evidence of current lab access or learner results. Exercises use synthetic data in a controlled, course-owned range. No real targets, reusable exploit code, malware or lure generation, or evasion are involved.
Source review
Primary-source review: October 31, 2026. References are linked below. Scenarios, scores, thresholds, and interface values are fictional course-design examples, not incident measurements or learner results. Modules and assessments remain planned until released.
The Agentic track's CVE identifiers, scores and fixed versions were checked against published CNA records on November 4, 2026. Other dated references retain their own review dates. Review the linked primary publications and current maintainer advisories before relying on a specification.
Planned and accessible modules
Planned means a curriculum preview exists, not that a lesson is available. Accessible means the owner has confirmed a released module and its direct Skool lesson destination. Skool sign-in and the relevant membership may be required. Joining Skool does not unlock planned modules. Planned labs, live-fire exercises and assessments require verification before release; setup requirements and any costs are stated with each module.
The proposed credential
Read the proposed sicurity.ai credential and assessment standards, including its unfinalized rules, verification and availability. A planned badge or sample record does not mean anyone has earned it.
Definition-specific limits
The glossary leads with definitions. These additional limits preserve the qualifications attached to those terms.
- Pathway
- A pathway listing is not evidence of released lessons or preparation for the whole role.
- Course
- Completing one course does not establish preparation for the whole AI Security Engineer role.
- Lab
- A preview is not current lab access.
- agentic AI
- Security work starts with its actual tools, identities, data access and stopping conditions, not with the label.
- AI gateway
- Passing requests through a gateway does not establish that their contents are safe or authorized.
- AI bill of materials (AI BOM)
- An inventory alone does not prove that a model is trustworthy, that licenses permit a use or that every dependency is secure.
- detection engineering for agents
- An unusual model response alone does not establish that an incident has occurred.
- evaluation (evals)
- Passing a test set is evidence about those tests, not proof that every future input or deployment will be safe.
- human verification
- A stated verification requirement is not evidence that a particular draft has already been verified.
- MITRE ATLAS
- A technique’s presence in the knowledge base does not establish that a particular product is vulnerable or that a training exercise reproduces a documented incident.
- model serialization attack
- A familiar filename or model name is not proof of a safe artifact.
- model supply chain
- A model’s published reputation does not replace those checks.
- NIST AI Risk Management Framework
- Using the framework is not a NIST certification, and citing it does not demonstrate that particular controls are effective.
- non-human identity
- A model’s instructions are not a substitute for identity controls.
- practice exam
- A practice score is not a guarantee of an outcome or a validated prediction of performance unless supporting evidence has been published.
- red teaming (AI)
- It is not permission to test other people’s systems or contact real targets.
- system prompt
- Sensitive credentials should not be placed in prompts on the assumption that instructions will keep them secret.
- track
- They do not establish current access or complete preparation for the AI Security Engineer role.
- trust boundary
- Crossing it should not silently turn untrusted content into instructions, permissions or evidence of a user’s intent.
Reviewed .