agentskills.codes
EM

embu-data-audit

by murzogluSource

Audit active EMBU data pipeline decisions in the doktoratezi repository. Use when working on EMBU, EMBU-C, EMBU-P, Likert 4pt/6pt, Stage 1, Stage 2, Stage 3, family matching, outliers, ICC, CFA, mixed Likert families, find_embu_columns, PII removal, or EMBU v2.0 validation.

Install

mkdir -p .claude/skills/embu-data-audit && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/15032" && unzip -o skill.zip -d .claude/skills/embu-data-audit && rm skill.zip

Installs to .claude/skills/embu-data-audit

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Audit active EMBU data pipeline decisions in the doktoratezi repository. Use when working on EMBU, EMBU-C, EMBU-P, Likert 4pt/6pt, Stage 1, Stage 2, Stage 3, family matching, outliers, ICC, CFA, mixed Likert families, find_embu_columns, PII removal, or EMBU v2.0 validation.
274 chars✓ has a “when” triggerlonger than Claude Code's old 250-char listing cap (fine on current versions)

About this skill

EMBU Data Audit

Use this skill for EMBU pipeline review, verification, or narrow fixes. The current project status is important: EMBU data architecture v2.0 is active. Previous inactive-status language was an obsolete artifact; CSV headers are treated as mislabeled, PDF is canonical, and the separate field-validation phase is no longer required.

Load First

Read these files before making claims or edits:

Then inspect the relevant implementation:

Audit Procedure

  1. State the audit scope: Stage 1 standardization, Stage 2 Likert conversion, Stage 3 family structure, or Stage 4 psychometrics.
  2. Confirm the active v2.0 decision context before interpreting CFA, ICC, or EMBU scale results.
  3. Check data-boundary safety: do not expose names, raw rows, credentials, or raw data contents.
  4. Verify column discovery through find_embu_columns() instead of positional assumptions.
  5. Verify Likert handling through classify_embu_c_likert() and mark_mixed_likert_families().
  6. Verify out-of-range values are converted to NA and logged to outputs/tables/embu_stage1_outliers.csv.
  7. Verify family structure uses aile_no x cocuk_no and that index/sibling roles are handled in Stage 3.
  8. Run the narrowest relevant tests, then broaden only if the changed surface requires it.

Verification Commands

Rscript tests/test_embu_stage1.R
Rscript tests/test_embu_stage2_likert4.R
Rscript tests/test_embu_stage3_family.R
Rscript scripts/R/01_embu_stage1_standardize.R
Rscript scripts/R/02_embu_stage2_likert4.R
Rscript scripts/R/03_embu_stage3_family.R

Use runner scripts only when the user asks for regeneration or when verification genuinely requires regenerated artifacts.

Response Shape

Return:

  • Scope audited.
  • Files inspected.
  • Findings ordered by severity.
  • Whether any active v2.0 decision caveats affect interpretation.
  • Tests or commands run.
  • Any residual uncertainty requiring additional documentation, code, or field-record review.

Search skills

Search the agent skills registry