TRACK 03 / OSWORK

Continuity

Save decisions and retrieve versions.

0% 0 of 0
01 / OSWORKAGENTS, Skills and memory02 / OSWORKGit and GitHub without losingwork03 / OSWORKPRACTICE04 / OSWORKCHECK
Read, apply to a small task and check the result before moving on.
2 modules
12 topics
85 min with practice
Guided step by step

Track map

3.1 ~40 min

🧭 AGENTS, Skills and memory

Create project instructions and a reusable capability with a review criterion.

3.2 ~45 min

🧭 Git and GitHub without losing work

Save a version, inspect differences and recover a training change.

Detailed content

MODULE 3.1

AGENTS, Skills and memory

Create project instructions and a reusable capability with a review criterion.

0% 0 of 0

What it is

AGENTS.md is the instruction file that Codex discovers in the applicable scope. It describes how to work: initial files, verification commands, limits and delivery format. No need to explain the whole organization history; prefer short rules that change a real decision.

Why learn

Objective instructions avoid repeating the same details in each conversation. The file should help the agent choose a concrete action, such as verifying the report before considering it finished.

Key concepts

Operational instruction; scope; observable rule; conciseness.

What it is

By default, ~/.codex/AGENTS.md stores global instructions. In the project, AGENTS.md adds specific rules; files in closer folders can prevail in the corresponding scope. AGENTS.override.md has priority over AGENTS.md at the same level. This does not override system instructions, permissions or the explicit request of the session.

Why learn

Path matters. A local rule may not apply to another folder, and a forgotten override file can explain unexpected behavior. Keep the global small and leave local details in the project.

Key concepts

Discovery; hierarchy; directory scope; override.

What it is

A Skill gathers instructions for a recurring activity, with name and description at the beginning of SKILL.md. It may include supporting resources and programs. Personal skills reside in ~/.agents/skills; project skills can be in .agents/skills within the repository. The ~/.codex folder remains Codex configuration.

Why learn

A rule says what to respect; a Skill teaches a procedure that can be triggered when needed. Separating these roles avoids loading the entire manual into every task.

Key concepts

Name; trigger description; procedure; input and output; validation.

What it is

The course's operational memory is a set of queryable files, not a change to the model weights. It works when the agent reads the relevant information and when someone keeps that information up to date. Store stable facts, decisions and causes of failures; do not preserve every sentence of every conversation.

Why learn

Copying the entire history increases volume and can reintroduce old instructions. A small, dated and reviewed memory helps more than a massive file that no one can validate.

Key concepts

External memory; explicit query; summary; validity; source.

What it is

Record the symptom, the observed cause, the minimal fix and how to verify. Differentiate failure of request, such as ambiguous objective, from infrastructure failure, such as a terminated process. The record only generates operational learning when it changes the next procedure.

Why learn

Redoing the whole project can mask a simple problem. A small protection, like checking for a file's existence before reading, is usually easier to test and maintain.

Key concepts

Symptom is not cause; minimal correction; prevention; evidence.

What it is

Test the Skill with a normal input, another incomplete one, and one out of scope. Observe whether the result preserves data, signals uncertainty, and stops when it should. The test should measure behavior, not just look for the Skill name in the response.

Why learn

A procedure that works once may be relying on accidental context. Varying inputs helps discover what needs to be made explicit in the instructions.

Key concepts

Normal case; incomplete case; scope limit; acceptance criterion.

View full →

MODULE 3.2

Git and GitHub without losing work

Save a version, inspect differences and recover a training change.

0% 0 of 0

What it is

Git records file versions. A repository is the folder tracked by this history; a commit is a record with changes and a message. GitHub is a service that hosts remote repositories. You can use Git locally without publishing anything on the internet.

Why learn

When an agent changes many files, the history allows you to understand what changed and recover a known version. Git does not replace a full backup: ignored files, databases, and external data need their own protection.

Key concepts

Repository; commit; history; remote; backup.

What it is

git status shows new, modified, and staged files. git diff shows changes not yet staged; git diff --cached shows what will go into the next commit. The staging area, called staging, lets you choose exactly which files belong to the same change.

Why learn

git add . stages everything that is not ignored. For learning, prefer naming files: you see the scope better and reduce the risk of including unrelated material.

Key concepts

Working tree; staging; diff; content review.

What it is

Configure user.name and user.email locally to identify authorship. Stage the desired files and use git commit -m with a concrete description. A commit should represent a change you can explain and verify.

Why learn

Messages like “update” make the history less useful. A version is only a reliable point if you know whether it worked and what checks were performed.

Key concepts

Authorship; message; cohesive change; verification.

What it is

git clone copies a remote repository and its history. git pull fetches and integrates changes into the current branch. Before updating, check git status. In an initial flow, git pull --ff-only only accepts a direct fast‑forward update and stops when histories have diverged.

Why learn

Updating a folder with local changes can cause conflicts. The --ff-only lock is useful information: do not bypass it by discarding work. Inspect the history or ask for help with the context.

Key concepts

Clone creates the folder; pull updates; branch is a line of work; divergence requires review.

What it is

git revert creates a new commit that undoes a previous change. It is appropriate for fixing a record that has already been shared. git restore discards unsaved changes of selected files; it can lose work. Do not teach reset --hard as an automatic response to any difficulty.

Why learn

Recovery tools have different consequences. Identify whether the change is only in the file, in a local commit, or published before choosing the command.

Key concepts

Revert preserves history; restore discards selected changes; recovery requires diagnosis.

What it is

git push sends commits to the remote. Before that, check the account, the destination URL, the scope of the files and the absence of credentials. A public repository remains accessible to third parties; a private one also requires access control. Publishing a site is an additional step, depending on the hosting.

Why learn

Mixing save and publish leads to accidental exposure. Separate “record locally”, “push to GitHub” and “put the site live” in your checklist.

Key concepts

origin; push; visibility; credentials; publication.

View full →
Full module