A Writer's Git: Minimal Workflow Covering 80% of Document-Engineering Needs with 20% of Git Surface

Document Engineering Vol 1 introduces 14-andup Vietnamese learners to version control as the second programmable substrate they encounter after the filename and folder convention. The volume requires a Git workflow specification that is both pedagogically defensible and operationally narrow: a learner must reach the end of the chapter able to track their own writing, recover from accidental edits, branch for trialrevisions, and synchronise work between a personal laptop and a remote like GitHub or GitLab, without ever encountering a merge conflict. The "no merge conflicts in Vol 1" constraint is not an aesthetic preference; it is a load-management decision derived from the cognitive-dimensions analysis of Git [1], the conceptual-misfits analysis [2], [3], and the empirical record of novice Git struggles in computing education [4], [5], [6]. The contribution is a fifteen-command writer's Git workflow specification covering initialisation, status, staging, committing, history inspection, single-file recovery, soft reset, branching, branch switching, push, pull (in the conflict-free configuration only), clone, ignore-file management, tagging, and identity-andsignature setup. The non-coverage inventory — merge with conflicts, rebase, cherry-pick, stash, reflog, submodules, forcepush, hooks, bisect, advanced reset modes, the index as a directlymanipulated object — is explicit, justified, and deferred to a Vol 2 collaboration chapter whose research basis is H2.3's sibling-node downstream. Four example session walkthroughs (solo writing, draft snapshot, accidental-revert recovery, trial-revision branch) and a Windows/macOS/Linux reproducible-setup section operationalise the specification. The synthesis draws on Chacon and Straub's Pro Git [7], Loeliger and McCullough's Version Control with Git [8], Demaree's Git for Humans [9], Lynn's Git Magic [10], Bryan's Happy Git with R [11], the Software Carpentry git-novice lesson [12], the official giteveryday(7) minimum-set documentation [13], and the empirical and conceptual literature listed above. No cohort work is conducted; audit-loop review is the quality bar per program scope.