enriquerodrig

citation-validator

Use this skill when modifying src/regulaitor/citation/validator.py or its policy. Documents the canonical 3-check validation procedure and the rules for evolving it (e.g. adding a fuzzy-fallback layer in H15).

Data & Documents Listed

document-analysis

Use this skill when extracting, sanitizing, segmenting, or analyzing a document end-to-end through the RegulAItor pipeline (PDF or Markdown). Activates the full extract→sanitize→segment→loop[gate→retriever→analyst→auditor]→aggregate flow with SSDLC-aligned defaults.

evals-runner

Use this skill when running the H8 evaluation harness, reading `evals/reports/latest.md`, deciding whether to re-run, or extending the gold set. Activates from H8 onwards.

prompt-versioning

Use this skill when adding, modifying, or rolling back agent prompts in src/regulaitor/agents/prompts/ to keep the project's prompt history reproducible and auditable. Activates from H4 onwards (Analyst, Auditor, Council).

rag-ingest

Use this skill when adding a new regulatory corpus (NIS2, DORA, or any future norma) following the H1 RegulAItor pattern. Ensures the new corpus integrates with the existing fetch/parse/validate/manifest pipeline without ad-hoc divergence.

redteam-runner

Use this skill when running the H9 red team suite, reading `redteam/reports/latest.md`, deciding when to re-run, or extending the attack set. Activates from H9 onwards.

secure-coding-checklist

Use this skill before merging any PR that touches src/regulaitor/security/, src/regulaitor/document/sanitizer.py, src/regulaitor/citation/validator.py, or src/regulaitor/agents/auditor.py. Activates from H9 onwards (CLAUDE.md §12.3.10).