Deep-analyze governed agent identity books

[murraylovecode]

Summary

• deeply analyzes the exact first editions of Solving the Bottom Turtle and Building Secure and Reliable Systems
• adds complete-edition access and reproducibility records, chapter and argument maps, and two original 26-section analysis pages
• stages 36 locator-backed source statements and 16 proposition candidates for governed AI-agent identity research
• records chapter-level attribution, related-source relationships, six later-author sources, and eight counterargument or qualifying sources
• adds a complete review worksheet, a 15-statement independent sample, rights review, limitations, and batch validation report
• adds a reusable validate:book-analysis command

Access and rights

• SPIFFE Project first-edition PDF: 194 pages, ISBN 978-0-578-77737-5, CC BY 4.0, SHA-256 recorded
• O’Reilly/Google first online edition: all 35 official HTML files, 176,914 normalized words, stable chapter/section anchors, manifest and normalized-text hashes recorded
• no third-party full text is committed
• no direct quotations are used

Review

• 18 statements per book, within the required 8–20 range
• 15/36 statements (41.7%) independently replayed, including all high-impact, criticism, and change claims
• four statements were narrowed or relinked after independent review, then passed replay
• all propositions remain single-source, staging-only candidates with ownership-concentration warnings
• no human approval or production promotion is implied

Validation

• batch-specific schema, access, locator, quote, duplicate, proposition-overlap, provenance, rights, review-coverage, near-duplicate prose, and staging-leak checks pass
• canonical repository validators and audits pass
• lint and Astro/TypeScript checks pass with zero errors or warnings
• 24 tests pass across four test files
• production static build and post-build generation pass
• git diff --check passes

Stack

This draft PR targets research/books-governed-agent-identities-BATCH-2026-002, the metadata batch branch, because the deep-analysis records extend the exact work and edition IDs introduced there.