project

What it does

project mode reads the project idea description, your target archetypes from config/profile.yml, and your existing portfolio context, then scores the project on three dimensions: signal value for target roles, differentiation from your existing portfolio, and estimated time-to-shippable.

When to use it

Use project when you have three or four side-project ideas and want to pick the one that actually moves your job search forward. It is the side-project equivalent of the oferta mode.

Example

`/career-ops project` + project description — output is a scored evaluation with a build/skip recommendation.

What a run looks like

A project run returns a legitimacy read, a weighted score across six dimensions, and one clear verdict. You hand the mode an idea, and it grades that idea the way a hiring manager would read it off your portfolio rather than judging whether the code will compile.

Portfolio Project Evaluation
URL: (idea only, no repo yet)
Legitimacy: High Confidence

Dimension                 Weight  Score
Signal for target roles     25%     5   Directly demonstrates a JD skill
Uniqueness                  20%     3
Demo ability                20%     5   Live demo in 2 min
Metrics potential           15%     4   latency, cost, accuracy
Time to MVP                 10%     4   ~1 week
STAR story potential        10%     4   trade-offs worth narrating

Verdict: BUILD
  Week 1 -> MVP with the core metric
  Week 2 -> polish + interview pack

The verdict is always one of three: BUILD, with weekly milestones; SKIP, with what to do instead; or PIVOT TO an alternative that carries more signal for the roles you are chasing. The six dimensions are weighted, so a project that demos well and produces clean metrics can outrank a more novel idea that has no visible payoff.

The interview pack

Every approved project ships with an interview pack, because the story is the deliverable, not the repository. The pack has three parts: a one-pager covering product, architecture, metrics, and evaluation plan; a demo, either a live URL or a two-minute recorded walkthrough; and a postmortem of what worked, what did not, and how you mitigated it. A built project with no narrative attached rarely moves a job search, while a modest project with a crisp trade-off story often does. The 80/20 plan reflects this: week one is the MVP with its core metric, week two is polish plus the pack itself.

Gotchas

The mode does not validate technical feasibility, only career signal value. Run a separate technical scoping pass.

Related

• training