Methodology and AI Disclosure
How Waspada Digital works.
Waspada Digital monitors official Malaysian and authoritative technical sources, groups related evidence into incidents, and publishes only when deterministic evidence gates are satisfied.
Updated: 29 July 2026
Important: There is no human editor. Ambiguous, accusatory or high-risk material remains unpublished. AI may structure supported facts but cannot create evidence or practical advice.
1. Collection
The system records source URLs, retrieval times, publication dates, snapshots, fingerprints and extracted claims from selected official sources. Collection does not mean an item will be published.
2. Grouping and evidence status
Notices about the same event or vulnerability are grouped into one incident. Each incident receives an explicit state: officially confirmed, technically confirmed, corroborated, developing, unverified, resolved or corrected.
Official confirmation requires a relevant Malaysian authority, regulator or affected organisation. Technical confirmation requires an authoritative vulnerability record plus an official vendor advisory.
3. Deterministic publication gate
Only records with adequate source provenance, safe content and sufficient context may become public. Unverified submissions, unsupported allegations and exposed personal identifiers are held. Downvotes and an absent lookup result are never evidence.
4. Structured explanation
A public alert explains what happened, confirmation status, why it matters, who is affected, what to do, what remains unknown, its supporting sources, when it was checked and what changed later.
Practical steps must come from authorities, affected organisations or official vendor documentation. AI may summarise and organise those steps but must not invent them.
5. Sources and attribution
Every alert links to its evidence records. “Verified” appears only for officially confirmed and technically confirmed records. Active exploitation requires qualifying evidence such as MyCERT, NACSA, CISA KEV or official vendor confirmation.
6. Corrections
Material updates, status changes, corrections and resolutions are recorded in the alert timeline. If an alert appears inaccurate, use the corrections process.