One-line verdict. The strongest comprehension in the cohort. The single criticism that defined her first evaluation — zero self-initiated work — has flipped completely from Week 5 onward.
3.69/5
Strong
Period-average weighted score across all 7 pillars. Scale: 4.5+ Outstanding · 3.5+ Strong · 2.5+ Competent.
How this is scored
This mirrors the program’s evaluation framework (see the site Evaluation section): seven
weighted pillars, each rated 1–5, plus the prompt-quality metrics and L1–L5 maturity levels the
trainers track weekly. Prompt Engineering carries the heaviest weight — in an AI-first program, how you
communicate with AI determines everything else.
Pillar weights
Pillar
Weight
Prompt Engineering
30%
Problem Solving
20%
Code Quality
15%
Security
10%
Accessibility
10%
Professional Skills
10%
Leadership
5%
Score scale
Score
Level
Meaning
5
Exceptional
can teach, innovates
4
Independent
solo, understands why
3
Guided
needs help, gets basics
2
Awareness
knows concept, can't do
1
Absent
nothing
Trajectory at a glance
Prompt quality score
Rises steadily from 22% (W1) to 54% (W7). Graduation target is 60% first-pass-quality.
Prompt maturity level
L1 vague → L5 strategic. Target: L3 by W8. Hansa reaches L4 within the period.
Security & accessibility in prompts
Both climb late as real backends and the shared app force the issue. Program target is 50%+ by week 12.
Pillar radar — W1 vs W7
The shape expands outward over the seven weeks — the model the trainers overlay monthly.
Weekly prompt tracking
Week
Prompt Score
First-Pass
Security
A11y
Level
W1Sep 1–7
22%
10%
0%
0%
L1
W2Sep 8–14
36%
35%
5%
10%
L2
W3Sep 15–21
40%
35%
5%
10%
L2
W4Sep 22–28
44%
40%
30%
10%
L3
W5Sep 29–Oct 5
52%
45%
35%
10%
L4
W6Oct 6–12
50%
48%
30%
40%
L4
W7Oct 13–19
54%
52%
45%
45%
L4
Period scorecard (average)
Pillar
Score /5
Weight
Weighted
Prompt Engineering
4/5
30%
1.20
Problem Solving
4/5
20%
0.80
Code Quality
4/5
15%
0.60
Security
3/5
10%
0.30
Accessibility
2/5
10%
0.20
Professional Skills
4/5
10%
0.40
Leadership
4/5
5%
0.20
TOTAL
100%
3.70/5
Weighted total above uses rounded period-average pillar scores; the headline 3.69/5 uses the exact averages.
Week 1 · Sep 1–7
Foundations & tooling
Environment setup, with two genuinely good steering files.
Mostly setup, but one own-idea spark: on 09-02 she drafted a full 5-element brief for a fitness/gym site (it never became a maintained project). Lots of conceptual curiosity about how web dev works end to end.
“if I can’t explain it, I don’t keep it (steering rule she wrote)”— 09-03
Prompt-rubric read
Prompts are a mix of two structured steering files (strong) and casual, typo-heavy one-liners (axes 1–4 low).
Did well
Her steering files already encode the program’s core values (prompt structure + explainability).
Watch / fix
Shaky git mental model; heavy reliance on “push/pull” muscle memory.
The typo pattern that persists all period starts here.
Week 2 · Sep 8–14
OpenEyes & strong semantics
Top-of-cohort execution on an assigned landing project.
3.55/5
Strong
▲ improving
36%
Prompt score
35%
First-pass
5%
Security
10%
A11y
L2
Maturity
Pillar scores — W2
Pillar radar — W2
What was built
OpenEyes art-center landing: letter-tile logo (her own design idea), facilities slideshow, themed footer, adults/60+ sections.
A real <table> timetable with <details> dropdowns, a staff panel, a kids’ gallery, and a CSS-dedup refactor.
Skills demonstrated
Strong semantic HTML (tables with caption/colspan, figure, address), CSS variables/BEM, JS-free interactivity.
Initiative & origin assigned
Still ~100% assignment-driven on the what. The first eval’s verdict held: after seven sessions, one project (OpenEyes) and zero self-initiated work.
“check the prompts I’ve uploaded today and let me know what I can change to make them way better”— 09-14
Prompt-rubric read
Her OpenEyes prompts use the full Context/Task/Constraints/Format/Edge-cases template cleanly (axes 1–5 strong), but aimed at an assigned task — so initiative, not prompt skill, is the cap.
Did well
Correct semantics under complexity; <details> reused for the gallery — skill compounding.
Maintainability instinct — asked to deduplicate CSS with a no-regression constraint.
Meta-learning — asked for a ranked critique of her own prompts.
Watch / fix
Broken #about anchor; no prefers-reduced-motion fallback.
Recurring typos; slow decisions (slideshow cycled through options).
Week 3 · Sep 15–21
Breadth & JavaScript explosion
Three new multi-page sites and the arrival of real JavaScript.
Euro-Travels — rotating 3D globe with planes, interactive country map, 6 themes, a site-wide navbar translator.
Rom-Project — Romanian-course site: hero video, a typewriter cycling “Romanian” across 5 languages, mega-menu, mobile nav.
Skills demonstrated
JavaScript clearly appears (translator, typewriter, globe, maps) — beyond the taught curriculum. Plus responsive, animation, SVG, video backgrounds.
Initiative & origin mixed
A big jump in ambition and JS skill, but these still read as assigned-style niche builds executed with unusually high flair. The “whose idea” question is improving but not yet settled.
“nope its not the way i wanted it — lets go with option A”— 09-17
Prompt-rubric read
Specificity and context (axes 1–2) high; iteration efficiency (axis 9) hurt by indecision. Security/a11y (7–8) still largely absent.
Did well
JavaScript going-beyond, across three rich sites; strong aesthetic direction.
Responsive + animation craft is well above beginner level.
Very typo-heavy; sometimes 15+ bullet edits in a single prompt.
Week 4 · Sep 22–28
Full-stack breakthrough
First full-stack work — PHP, MySQL, JWT, Angular.
3.80/5
Strong
▲ improving
44%
Prompt score
40%
First-pass
30%
Security
10%
A11y
L3
Maturity
Pillar scores — W4
Pillar radar — W4
What was built
Pivoted the blog into a real PHP + MySQL + MAMP app: login with prepared statements + password_verify(), “don’t reveal whether email or password was wrong.”
An Angular frontend: login service, posts list, admin table with edit/delete modals, categories, search, image upload, a WYSIWYG editor.
Skills demonstrated
PHP, MySQL, JWT auth, Angular (components/services/routing), REST, Postman, MAMP, file uploads — a major technical escalation.
Initiative & origin assigned
Still assignment-flavored on the what, but independently driven. Sharp conceptual questions and strong metacognition — she asked the assistant to flag “this is new to you” so she’d know where she has more to learn.
“issue a signed JWT containing the user id … a reusable check that protects my other endpoints by verifying that token.”— 09-25
“if I want a UI backed by a database, how will that work?”— 09-27
Prompt-rubric read
Security (axis 7) strong and partly self-driven; her backend prompts are cleaner and more precise than her W3 CSS walls.
★ Turning point — she originates two projects of her own and runs both documentation-first.
4.45/5
Strong
▲ improving
★ Turning point. This is the week the “zero self-initiated work” story changes — she originates two projects of her own and runs both documentation-first.
52%
Prompt score
45%
First-pass
35%
Security
10%
A11y
L4
Maturity
Pillar scores — W5
Pillar radar — W5
What was built
Valhalla — a dark/premium gaming shop (offline-payment reservation flow, blog, admin dashboard), spec’d across 10 phases.
TaskFlow — a second self-initiated task manager (status workflow, archive-never-delete, per-team task IDs, guest access with expiry).
Skills demonstrated
Software architecture/requirements thinking, data modeling, API design, auth/security design, NestJS/TypeORM/Angular stack selection.
Initiative & origin own work
Maximal. Two projects, both her idea, both method-driven. She chose the stack herself and reasoned about HttpOnly-cookie vs. localStorage tradeoffs and bcrypt. The first eval’s single blocker — “a project with her name on the idea” — is resolved here.
“we’re putting all the creativity into this project … we only going to make the documentation first … when it comes to questions, don’t hold back at all.”— 09-29
Prompt-rubric read
Her strongest prompt-engineering of the whole period — purposeful, tradeoff-aware, self-authored. Problem-solving and leadership peak here.
Did well
Two self-conceived projects run as proper specs — senior-beginner behaviour.
Reasons about security and architecture tradeoffs on her own (HttpOnly cookies, bcrypt).
Reaches L4 maturity: security + edge cases baked into the planning.
Watch / fix
Defers some final tech picks to the assistant’s recommendation.
Verbose; typos continue.
Week 6 · Oct 6–12
Delivery & joining the team
Ships the Valhalla build and plugs into the shared repo.
4.00/5
Strong
▬ steady
50%
Prompt score
48%
First-pass
30%
Security
40%
A11y
L4
Maturity
Pillar scores — W6
Pillar radar — W6
What was built
Took the Valhalla docs into a real NestJS + Angular + MySQL build (admin panel, image upload, design tokens, a fluent mobile-first refactor).
Ran the shared TaskBoard locally and asked a sharp data-safety question before touching it.
Began contributing to shared Landaro — Legal module data model, filling the “dependencies” sections without changing files.
Skills demonstrated
NestJS/Angular/MySQL full build, SCSS/design tokens, mobile-first responsive, reading and running a teammate’s project, cross-module documentation.
Initiative & origin own work
High and now collaborative — both delivering her own project and plugging into a real multi-person codebase.
“read the data model — ask to understand it, don’t touch the files”— 10-05
“Ask to learn, don’t touch the files” is exactly the right instinct for a shared repo.
Accessibility rises with a deliberate mobile-first refactor.
Watch / fix
Mostly steady execution — no major gaps this week.
A single “pull from git” day (10-12) keeps volume uneven.
Week 7 · Oct 13–19
Team member + second own build
Confident team contribution plus a second self-initiated full build.
4.50/5
Outstanding
▲ improving
54%
Prompt score
52%
First-pass
45%
Security
45%
A11y
L4
Maturity
Pillar scores — W7
Pillar radar — W7
What was built
Landaro (shared) — reorderable admin Legal categories persisted server-side, “Learn More” FAQ overlay, a consultant showcase, and a full course study-tracker (progress rings, role-gated, wired to real backend data).
She owned team git hygiene herself — per-feature branches, a database-naming scheme, and a before-pull clash check.
Dino-Lab (her own) — a new self-conceived dinosaur learning site + shop + blog, full build (fresh ng new, isolated PHP API port, its own DB, section-by-section with live verification).
Skills demonstrated
Real team git workflow (branches, clash avoidance, cross-OS), Angular + PHP/MySQL feature delivery against an existing backend, SCSS partials with @use and mixins, role-based UI gating, a repeatable spec→build process.
Initiative & origin own work
Highest of the whole period — a second self-initiated full build plus substantive shared-project contribution, with her proactively owning the team’s data-clash safety process.
“there are 3 people working … I want to name my database (ex: database-hansa) … before pull every time, check if there’s anything new in the remote that might clash, and let me know the solution before executing it.”— 10-16
Prompt-rubric read
The heavy specs were front-loaded in planning docs, so per-turn prompts look sparse — but the quality lives in the briefs. Leadership and problem-solving at their highest.
Did well
Designed the team’s data-clash safety process herself — leadership in a shared codebase.
Second self-initiated full build (Dino-Lab) using a repeatable spec→build method.
Code quality and problem-solving peak: role-gated UI wired to real backend data, SCSS architecture.
Watch / fix
Many single-word “go” prompts make per-turn prompt quality look thin out of context.
The W2 craft gaps (broken #about, no reduced-motion) were never explicitly revisited.
DevSchool · Summary
Hansa Herath — overall
Seven-week trajectory, strengths, and priorities for the next period
Week-by-week summary
Week
Focus
Weighted
Grade
Level
Origin
W1
Foundations & tooling
2.05
Developing
L1
assigned
W2
OpenEyes & strong semantics
3.55
Strong
L2
assigned
W3
Breadth & JavaScript explosion
3.45
Competent
L2
mixed
W4
Full-stack breakthrough
3.80
Strong
L3
assigned
W5
Her own projects, her own method
4.45
Strong
L4
own work
W6
Delivery & joining the team
4.00
Strong
L4
own work
W7
Team member + second own build
4.50
Outstanding
L4
own work
Strengths & priorities
Clear strengths
Comprehension-first habit — constantly asks for concepts to be explained and for critiques of her own prompts. Strongest understanding in the cohort.
Security instincts — hashing, JWT, HttpOnly cookies, “don’t reveal which field was wrong.”
Spec-before-code discipline — documentation-first on both of her own projects.
Initiative — resolved — Valhalla, TaskFlow and Dino-Lab are all her own ideas (W5 onward).
Real collaboration — she designed the team’s data-clash safety process herself.
Priorities for next period
Precision & typos — persistent across all seven weeks (“sutaible”, “intractive”, “peroper”). Cosmetic but constant; tighten before shipping.
Decide faster — a tendency to over-specify then reverse (the W3 globe/map saga, the W2 slideshow).
Own the final call — reduce reliance on “go with the recommended option”; make and defend her own technical decisions.
Close old craft gaps — the W2 broken #about anchor and missing prefers-reduced-motion were never confirmed fixed.
Environment fluency — recurring PowerShell/MAMP/port friction; worth drilling until it stops costing her time.
Bottom line. The defining criticism of her first evaluation — the widest gap in the cohort between ability and initiative — is no longer true after Sep 29. She now originates her own projects and contributes to a shared production-style repo with professional git and security practices. Coverage, understanding and going-beyond are top-tier; initiative is now strong. What caps her is polish — typos, decision churn, and deferring final calls — not capability.
own work self-authored idea/prompt
mixed partly assigned, partly own
assigned executing an instructor brief