The Agent Skills Directory

[SAFE]: The skill is entirely instructional, providing architectural principles and 'rituals' for documentation management based on industry-standard methodologies.
[SAFE]: The recommendations for 'reconciliation checks' and 'enforcement' (such as pre-commit hooks or linting) are presented as conceptual workflows for the developer or agent to implement using standard tools, rather than providing any scripts that execute automatically.
[SAFE]: No prompt injection, data exfiltration patterns, or hardcoded credentials were detected in any of the provided files.
[SAFE]: External references are limited to established technical literature (e.g., books by Cyrille Martraire and Martin Fowler) and industry concepts (ADRs, RFCs).
[SAFE]: The skill uses a 'references/' directory structure to store deep-dive content, which is a standard pattern for organizing agent skills without overloading the primary context.

living-documentation