MODULE 2.2 / 4 OF 8

Folders, Markdown and secrets

Build the digital house and separate knowledge from credentials.

0% 0 of 0
01 / OSWORKA folder represents a context02 / OSWORKMarkdown is organized text03 / OSWORKEach file has a job04 / OSWORKSecrets are not shareableknowledge.05 / OSWORKIgnore before the first commit.06 / OSWORKDo a context cleanup.
The six pieces of this module. At the end, you produce a learning evidence.
6 topics
~35 min reading and practice
1 verifiable delivery
Guided training lab

1A folder represents a context

What it is

The ~ symbol represents the home folder in Bash. Inside it, projects gathers independent work. Use clear names and avoid mixing documents from different clients. The config folder stores cross‑cutting knowledge; each project keeps its own inputs and results.

Why learn

Separate contexts help limit what the AI needs to read. A folder full of unrelated topics adds noise and makes it hard to explain where a conclusion came from.

Key concepts

Personal folder; projects; context; inputs and outputs.

In practice

In ~/projetos/website the site files reside. In ~/projetos/estudos the experiments reside. On Windows, the manager may show paths like C:\Users\SeuNome\projetos.

✓ Do it

Draw the tree before creating files. Choose a single training project and a single global config folder.

✗ Avoid

Accepting a conclusion without checking the input that supports it.

2Markdown is organized text

What it is

Markdown uses simple symbols to organize text: # creates a title, ## creates a subtitle and a hyphen starts a list item. The file remains plain text, readable even without a special editor. The name ends with .md. You do not need to write code to record clear instructions.

Why learn

Small, named, easy‑to‑edit files last longer than a lost conversation. They can be reviewed by people and consulted by agents. The value comes from clarity, not elaborate formatting.

Key concepts

Title; list; code block; link; plain text.

In practice

A README can contain: purpose, input files, expected result and how to verify. Anyone arriving later understands the task without relying on the original conversation.

3Each file has a job

What it is

memoria.md records stable preferences; decisoes.md explains choices; dicas.md stores useful procedures; falhas.md documents problems and fixes. Do not put everything in one giant document. When a decision changes, record the date and reason so you don’t keep contradictory rules.

Why learn

Separating functions makes it easier to consult only what’s needed. A failure history should not become a list of mandatory commands in every task. Queryable knowledge and permanent instructions are different things.

Key concepts

Selective memory; dated decisions; procedure; history.

In practice

"I prefer short reports" is a preference. "We chose CSV because it is compatible with the team's spreadsheet" is a decision. "The service stopped without supervision" belongs to failures.

Sequence to try

  1. Prepare a training copy.
  2. Distribute five fictional notes among the four files. For each one, explain why that is the appropriate place.
  3. Record the observed result and the next correction.

4Secrets are not shareable knowledge.

What it is

A .env file can store variables such as TELEGRAM_BOT_TOKEN or DATABASE_URL. It is not encrypted: anyone with access to the file can read it. Use appropriate permissions and never include real values in screenshots, course examples, or files sent to the AI without necessity.

Why learn

Credentials allow actions on behalf of an account. Separating the .env.example model, without real values, from the local .env lets you share the structure without distributing access.

Key concepts

Variable; secret; .env.example; runtime reading.

In practice

The kit includes TELEGRAM_BOT_TOKEN=fill_locally. The student replaces this only in their private copy. No bot is authenticated with this example.

✓ Do it

Create .env.example with variable names and dummy values. Keep .env out of the repository and never paste your key in the chat.

✗ Avoid

Mix the training copy with private files or production work.

5Ignore before the first commit.

What it is

The .gitignore lists files that Git should ignore when they are not yet tracked. Include .env, private variants, and temporary folders. Keep an explicit exception for .env.example. Before committing a version, examine git status and the staged files.

Why learn

Ignoring later does not erase a secret from history. If the key leaked, the first fix is to revoke or rotate it at the source; deleting the line from the file does not invalidate a copy already seen.

Key concepts

Tracked files; exclusion patterns; change review; revocation.

In practice

Useful patterns: .env, .env.*, !.env.example, __pycache__/. To discover which rule applies, use git check-ignore -v .env.

6Do a context cleanup.

What it is

The AI does not automatically read every existing Markdown file on the computer. Specify which documents to consult and keep references in AGENTS.md when needed. Before a task, remove irrelevant data from the working copy and update expired decisions.

Why learn

Useful memory needs to be findable and correct. An old file can be more harmful than no memory if it contains a service address or process that has already changed.

Key concepts

Context selection; date; source of truth; periodic review.

In practice

When starting a report, request reading of README.md and the decision about format. Do not load contact lists or credentials because they are in the same folder.

Criteria to review your delivery

Use this rubric after the lab. Each line asks for evidence; checking reading does not mean the practice was performed.

Criterion Expected evidence If not passed
Scope The delivery matches the objective of this class. Reduce the task and name a single result.
Inputs You know which files or data were used. List the sources and remove unrelated material.
Execution The procedure was carried out in the training environment. Differentiate what was planned from what was done.
Verification A result was compared with a reference. Open the file or repeat the verifiable query.
Secrets No token, password, or private data was shared. Review the work copy before any submission.
Continuity Another person can find the next step. Update README and record a concrete pending item.

GET HANDS-ON / ~15 MIN OR IN STEPS

Organize your second operational brain

Use fictional files and a training folder. Practices involving installation, Telegram or VPS may require additional time for registration and configuration.

Work structure

Read the block before using. Fields like Your Name and usuario@ip-da-vps are examples to adapt; administrative commands belong only to your training environment.

~/projetos/
├── config/
│   ├── memoria.md
│   ├── falhas.md
│   ├── dicas.md
│   └── decisoes.md
└── meu-primeiro-projeto/
    ├── AGENTS.md
    ├── README.md
    ├── .gitignore
    ├── entradas/
    └── saidas/

Ready criterion

Build the digital house and separate knowledge from credentials. Record the produced file, the test run and the observed result.

Open lab files and models

Check what remained

Adding .env to .gitignore automatically removes a key that has already been published?

View commented answer

No. Revoke the exposed key and fix the history as appropriate; ignoring only prevents new untracked files.

If your answer was different, return to the corresponding topic and write the difference in one sentence. The check does not block your study.

Module summary

  • Personal folder; projects; context; inputs and outputs.
  • Title; list; code block; link; plain text.
  • Selective memory; dated decisions; procedure; history.
  • Variable; secret; .env.example; runtime reading.
  • Tracked files; exclusion patterns; change review; revocation.
  • Context selection; date; source of truth; periodic review.

Consult the source

Tools verified on 20/09/2026; screen names and availability may change.

Full module