Model Package

Pure business logic for deriving jobs, agent profiles, skill matrices, and career progression from schema data.

When to Use

•Modifying job derivation logic (skill levels, behaviour maturities)
•Changing agent profile generation
•Working with skill modifier resolution
•Updating checklist derivation for stage transitions
•Modifying interview question selection
•Changing career progression analysis

Package Structure

code

apps/model/src/
  derivation.js   # Core: deriveSkillMatrix, deriveBehaviourProfile, deriveJob
  modifiers.js    # Skill modifier resolution
  profile.js      # Post-processing: filtering, sorting for agents
  job.js          # Job preparation for display
  job-cache.js    # Job caching for pages
  agent.js        # Agent profile/skill generation
  checklist.js    # Stage transition checklists
  interview.js    # Interview question selection
  matching.js     # Job matching logic
  progression.js  # Career progression analysis

Key Modules

derivation.js

Core derivation functions used by all consumers.

javascript

import {
  deriveSkillMatrix,
  deriveBehaviourProfile,
  deriveJob,
} from "@forwardimpact/model/derivation";

profile.js

Agent-specific filtering and sorting:

•Excludes isHumanOnly skills
•Keeps only highest level per skill
•Sorts by level (highest first)

javascript

import { prepareAgentProfile } from "@forwardimpact/model/profile";

job.js / job-cache.js

Job preparation and caching for web pages.

javascript

import { getOrCreateJob } from "@forwardimpact/model/job-cache";

agent.js

Agent profile and skill file generation.

javascript

import {
  deriveAgentSkills,
  generateStageAgentProfile,
} from "@forwardimpact/model/agent";

interview.js

Question selection based on role requirements.

progression.js

Career progression analysis between grades.

Job Derivation

Jobs are derived from Discipline × Grade × Track?:

•Base skill levels from grade (by skill type: primary/secondary/broad)
•Track modifiers applied to capabilities (+1, 0, -1)
•Behaviour modifiers combined from discipline and track
•Responsibilities by discipline type (professional/management)

code

Final Level = clamp(Base Level + Track Modifier, awareness, grade ceiling)

Agent Profiles

Agent profiles use the same derivation as jobs, filtered for AI-applicable capabilities:

code

Agent Profile = Discipline × Track × Stage

•Uses practitioner-level reference grade
•Excludes isHumanOnly skills
•Constrained by stage tools and permissions

Pure Functions

All derivation functions are pure:

•No side effects
•Same inputs → same outputs
•No external state access

This makes them easy to test and reason about.

Verification

Run tests after changes:

npm run test