How this portal works

An application portal that maintains itself

Not a document that goes stale as it is written — a repository an agent keeps current on a two-week cycle.

The mechanism

Demonstrate progress — answer questions, show artefacts.
Assess honestly — on a 0 to 4 scale, without flattery. Unconfirmed stays unconfirmed.
Write the protocol — as a file in the repository, including the original answers.
Generate the website — from those same files, in both languages.

These rules are not good intentions in a concept document; they are checks in the build. If one is violated, no website is produced.

No level without evidence
A skill level with no evidence behind it is not shown. Measured levels and agent estimates are distinguished on every page.
Certificates are called what they are
Only a passed examination may be called a certificate. Certificates of attendance are listed as such — even where the stronger wording would look better.
Awards only with a source
A second award that exists in memory is not listed, because no evidence could be found for it. It returns if a certificate turns up.
No client names without release
Logos on your own company website are not a documented reference release. Industries are named here instead of clients.
No private data
Date of birth, address, marital status and current salary appear nowhere — not even behind the login. For the same reason there are no reference PDFs to download: the originals contain those details.
The result is not sugar-coated
The measured level of knowledge is public, including where it sits below target.

Every change to content or portal is a commit. This list is generated from the Git history at build time, not maintained by hand.

No Git history was available for this build.