Files
SSW590-DevOps/backmatter/devopsAIPromptLog.tex
T

161 lines
9.6 KiB
TeX

\chapter{AI Prompt Log}
\ChapterLabel{AIPromptLog}
\index{AI prompt log}
\index{AI assistance}
\section{Scope of Assistance}
This appendix records the \gls{ai} assistance used in preparing this document,
in the detail needed for a reviewer to see what was requested, what was
returned, and what the author changed. Assistance was used for interpreting the
assignment brief, drafting chapter text, and reviewing the document for internal
consistency. It was not used to generate command output, evidence, or any claim
about work that was not performed. No passwords, access tokens, private keys, or recovery codes were supplied.
The supplied coursework evidence includes author names, local usernames, host
names, and the local network details visible in the terminal captures.
Every decision recorded below was made by the authors; the assistant's role was
to propose and to check.
\section{Records}
\subsection{Record 1 --- Document Interpretation and Restructuring}
\textbf{Date:} September 17, 2026.\\
\textbf{Tool and model:} Anthropic Claude (Opus 5).\\
\textbf{Task:} Determine what was expected by this assignment given its
instructions and the document template; then identify which parts of the template
belong in a submitted report and trim/modify unnecessary sections.\\
\textbf{Inputs:} The assignment brief and the supplied \LaTeX\ template sources.\\
\textbf{Output used:} Removal of the instructional example chapter and the
template-maintenance appendices; rewritten abstract, introduction, and document
history; corrected document metadata and authorship.\\
\textbf{Human review and corrections:} The assistant explained assignment expectations
and received clarification from the author according to class and lecture content.
The assistant provided a guideline on which template sections should be kept, removed,
or modified. The author identified sections that were erroneously retained and modified
the AI outputs heavily in the final iteration to match the scope of the assignment and lectures so far.
\subsection{Record 2 --- Chapter Drafting}
\textbf{Date:} September 18, 2026.\\
\textbf{Tool and model:} Anthropic Claude (Opus 5).\\
\textbf{Task:} Review the configured Kanban board and draft \ChapterRef{KanbanSetup},
then draft \ChapterRef{Hosts} and \ChapterRef{Passwords} according to the author's plans.
Update the AI log once all changes are finalized.\\
\textbf{Inputs:} A screenshot of the board and its column settings; the
intended platform and naming convention, and the access model already decided
by the author. No passwords, nor their hints or rules, were shared with the assistant.\\
\textbf{Output used:} The Kanban chapter structure, its column policy, and card
inventory tables; the host inventory and its naming convention; the
split of accounts into generated and derived classes.\\
\textbf{Human review and corrections:} The author determined the column policy
and work-in-progress criteria, and rejected a proposal to create a GitHub
organization as premature while teams remain unformalized. Much of the assistant's
wording, structure, and ambitious ideas were rejected for overcomplication.
\subsection{Record 3 --- Linux Commands}
\textbf{Date:} September 26, 2026.\\
\textbf{Tool and model:} Anthropic Claude (Opus 5).\\
\textbf{Task}: Draft a template for \ChapterRef{LinuxCommands} with appropriate formatting
for terminal inputs and outputs, and a structure for the problem set solutions. Revise
the entire document to reflect the new chapter and team formation.\\
\textbf{Inputs:} The document template and the Linux problem set.\\
\textbf{Output used:} A skeleton of the chapter with its sections, setup figure, and
the problem set headings for each question. Suggested edits to the metadata, abstract, and
introduction for Assignment~2 and the two-person team.\\
\textbf{Human review and corrections:} The authors refined the chapter structure to accommodate
rubric requirements and experimented with different figure formatting for terminal inputs and outputs.
\subsection{Record 4 --- Guided Execution and Evidence Integration}
\textbf{Dates:} September 27, 2026.\\
\textbf{Tool:} OpenAI Codex.\\
\textbf{Task:} Guide Victor through the Linux setup and Q1--Q40, check supplied
results against the assignment, review the group chapter, and integrate evidence
and concise explanations while preserving Payton's existing work.\\
\textbf{Inputs:} Assignment and problem-set PDFs, the group PDF, supplied
chapter and supporting LaTeX files, Victor's screenshots, and his terminal
transcript.\\
\textbf{Output used:} Command guidance; evidence checks; revised environment
description; per-question explanations; labeled terminal screenshots; Q15/Q20
clarifications; explicit Q38 assumptions; history and version updates.\\
\textbf{Human contribution and review:} Victor executed the commands and supplied
the resulting screenshots and transcript. Payton supplied the existing setup
figures and problem-set answers. Codex edited the documentation and checked it
against the recorded results; it did not generate substitute execution output.
The team remains responsible for reviewing the additions and compiling the
combined Overleaf project.\\
\subsection{Record 5 --- AWS Deployment and Proposal Draft}
\textbf{Dates:} October 1--5, 2026.\\
\textbf{Tool:} OpenAI Codex.\\
\textbf{Task:} Guide Victor through AWS account access, EC2 provisioning,
class-based website development, Docker installation and deployment, then draft
his chapter from the actual sources and screenshots.\\
\textbf{Human contribution:} Victor operated AWS, created the application
files, executed local and remote commands, resolved the repository-key failure
with guidance, and provided the execution screenshots. Payton is responsible
for the separate TeX Live compiler exercise and chapter. Both authors will
review the tool-comparison proposal.\\
\textbf{Output used:} Command guidance, source code, saved evidence, deployment
chapter, draft GitHub Actions/Jenkins proposal, metadata, and history updates.
Codex copied additions into the student's supplied course repository location.\\
\textbf{Review boundary:} Proposal selection requires team review. Payton's
compiler work was not present in the inspected repository. Full compilation,
combined report review, repository publication, and Canvas submission are not
claimed by this record.
\subsection{Record 6 --- LaTeX Docker Chapter}
\textbf{Dates:} October 5, 2026.\\
\textbf{Tool and model:} Anthropic Claude (Opus 5.5).\\
\textbf{Task:} Guide Payton through building a Docker container that compiles
a \LaTeX\ document with TeX Live and draft \ChapterRef{LatexDocker};
propose matching updates to the abstract, \ChapterRef{Introduction},
\ChapterRef{Hosts}, glossary, metadata, and document history.\\
\textbf{Inputs:} The assignment brief, Payton's terminal output and screenshots,
and Overleaf's TeX Live documentation.\\
\textbf{Output used:} The finalized Dockerfile, \texttt{.dockerignore}, and sample document;
the chapter draft; proposed replacement text for outdated sections.\\
\textbf{Human decisions and corrections:} Payton selected Debian's TeX Live packages,
built and ran the container under WSL, supplied the evidence screenshots,
and revised the chapter's organization and wording.\\
\textbf{Review boundary:} The assistant compiled the full report with its own
TeX Live installation to check the chapter; the image was built and
run only by Payton. All execution evidence was produced by Payton. The assistant
edited no repository files.
\subsection{Record 7 --- Inviscid Project Proposal}
\textbf{Date:} October 6, 2026.\\
\textbf{Tool:} OpenAI Codex.\\
\textbf{Task:} Brainstorm Inviscid, refine the metrics-tool comparison and
research question, and write the approved proposal into the course repository.\\
\textbf{Inputs:} The assignment outline, instructor clarification, Victor's
project preferences and corrections, current repository sources, and official
Prometheus and VictoriaMetrics documentation.\\
\textbf{Output used:} Project description, sample task, tentative tool choices,
research question, planned comparison, independent SWOT assessments, and
supporting document updates.\\
\textbf{Human decisions and corrections:} Victor approved the monitoring-service
concept, selected VictoriaMetrics instead of InfluxDB, approved the research
question, and requested independent SWOT assessments and candidate tools
explicitly marked undecided.\\
\textbf{Review boundary:} The monitoring prototype, tool comparison, and
pipeline remain proposed work. No performance results or completed monitoring
features are claimed. The combined report still requires final author review
and compilation before submission.
\subsection{Record 8 --- Gitea Foundation and Evidence Draft}
\textbf{Dates:} October 12--15, 2026.\\
\textbf{Tool:} OpenAI Codex.\\
\textbf{Task:} Guide Victor through Student Pack application, Docker/SQLite
Gitea setup, SSH tunneling, Elastic IP association, and Payton's key access;
prepare the pre-PR documentation from supplied execution evidence.\\
\textbf{Human contribution:} Victor operated the consoles and terminals,
created accounts, supplied screenshots, and reported Payton's successful SSH
connection. Payton's remaining work was not executed by the assistant.\\
\textbf{Output used:} Non-secret Compose configuration, deployment notes,
saved screenshots, Gitea chapter draft, and matching inventory, account-hint,
history, and metadata updates.\\
\textbf{Corrections and limits:} Documented benefit activation separately from
academic approval; recorded the current Elastic IP and the unresolved campus
SSH timeout; did not infer public HTTPS or persistence replacement success
from local checks. Domain details, renewal and external checks, PR workflows,
final compilation, and submission remain unverified.