ms-shamanth/career-data-toolkit

v1.0.0MIT

Engineering knowledge for InternPilot AI career data: normalizing job and profile skills, computing and explaining the deterministic 0-100 match score, and comparing a resume against a specific job. Knowledge only; all application logic lives in the project code.

GitHub

career-data-toolkit

Custom Kiro Power for InternPilot AI. It packages engineering knowledge about career data (skill normalization, job normalization, match scoring and explanations, resume analysis) as three Agent Skills plus an exact copy of the scoring rules.

career-data-toolkit/
  plugin.json                         Agent Plugins 1.0.0 manifest (Kiro fields under extensions["dev.kiro"])
  POWER.md                            legacy-format overview and front matter
  skills/job-normalization/SKILL.md
  skills/resume-analysis/SKILL.md
  skills/match-explanation/SKILL.md
  references/scoring-rules.md         design.md §5.1-5.6, algorithm_version 1.0.0

Activation keywords: skill normalization, job normalization, match score, match explanation, resume analysis, internship, career data.
Used by backend-agent (.kiro/agents/backend-agent.json lists the skills as resources) and by any Kiro chat that mentions the keywords.
Knowledge only: the implementation lives in backend/app/services/matching/, backend/app/services/ingestion/ and backend/app/services/resume_service.py.
Why displayName and skills sit under extensions: the Agent Plugins 1.0.0 manifest schema is closed (only $schema, name, version, description, author, homepage, repository, license, keywords, extensions are allowed at the top level) and skills are discovered from the fixed skills/ folder. Client-specific data belongs under a reverse-domain namespace.

Full documentation: docs/kiro-powers.md.