Skip to main content

Documentation Complexity Targets v0.1

Last updated: 2026-07-15

Purpose: set fail-closed measurable complexity budgets for AI retrieval quality.

Metrics

  • active_docs_md: markdown files under docs/ excluding redirect stubs that carry the top-level deprecation marker.
  • active_top_level_leaf_docs: non-deprecated docs/*.md leaf docs excluding infrastructure docs (docs/INDEX.md, docs/DEPRECATION_MAP_v0.1.md).
  • capability_retrieval_fanout: primary_evidence_paths / capability_rows from feature_matrix.md.

Targets (v0.1)

  • active_docs_md <= 135
  • active_top_level_leaf_docs <= 6
  • capability_retrieval_fanout <= 0.46

Enforcement

  • Policy file: policies/docs/doc_complexity_budget.toml
  • Gate: scripts/check_doc_complexity_budget.sh
  • Read-only check: bash scripts/check_doc_complexity_budget.sh renders into a temporary directory.
  • Explicit local report update: bash scripts/update_doc_complexity_report.sh writes .genesis/perf/doc_complexity_report.json.

These targets are designed to reduce retrieval ambiguity for agent authoring while preserving required normative coverage.