NAVIGATOR
Assessments
COMPASS capability

Six-dimension capability baseline

Score each dimension from Novice to AI-native to produce a baseline you can re-assess against later. This sandbox isn't persisted yet — production wires it to ATLAS so deltas show up in delivery telemetry.

  • Prompting fluency

    Crafts prompts that elicit reliable, structured output.

    1. Treats LLMs as search2. One-shot prompts, no structure3. Iterates with examples4. Uses tool/function patterns5. Designs reusable prompt assets
  • Agentic workflow design

    Composes multi-step agentic flows with judgment about handoffs.

    1. Manual everything2. Single-agent calls3. Sequenced calls4. Multi-agent with handoffs5. Designs flows others use
  • COMPASS methodology grasp

    Knows the seven phases, three fidelity levels, six transformations.

    1. Unfamiliar2. Heard the terms3. Names the phases4. Runs phases live5. Mentors others through them
  • Artifact literacy

    Reads and produces COMPASS artifacts at the right fidelity.

    1. Skips artifacts2. Generates without review3. Reviews own output4. Drives artifact chains5. Sets quality bars for team
  • Evaluation rigor

    Critically evaluates AI-generated work; refuses bad output.

    1. Accepts as-is2. Spot checks3. Structured review4. Rubric-driven5. Defines team rubrics
  • Collaboration cadence

    Operates handoffs, decisions, and reviews on shared infrastructure.

    1. Solo, ad-hoc2. Inconsistent handoffs3. Uses ATLAS sessions4. Drives clean handoffs5. Designs team operating model
Overall · individual
3.0/ 5.0

Level: Practicing

You're ready for the full Practitioner track. Pair it with hands-on COMPASS sessions in ATLAS.