Skip to content
← Courses

Level 5

Skills, hooks, permissions and security

Extend Claude Code with skills, hooks and plugins, and secure its use: permissions, sandboxing, prompt injection, secrets and review.

6 modules 35 tasks ≈ 25 h of work

What you will be able to do

  • Create skills that trigger at the right time, with scripts and resources
  • Automate checks with hooks on Claude Code events
  • Package and distribute extensions as plugins
  • Apply least privilege with fine-grained permissions, managed settings and sandboxing
  • Identify and mitigate prompt injection and secret exposure
  • Set up a systematic, security-focused review of generated code

Level content

  1. Module 1

    Agent Skills

    Package know-how into skills that Claude loads only when they are useful.

    1. What is a skill? Instructions loaded on demand30 min
    2. Skill structure: SKILL.md, frontmatter and supporting files40 min
    3. Writing a description that triggers the skill at the right time35 min
    4. Adding scripts and resources to a skill45 min
    5. Personal skills, project skills and shared skills30 min
    6. Workshop: creating a skill for one of your team's processes75 min

  2. Module 2

    Hooks

    Automatically run your own commands at specific points in the Claude Code lifecycle.

    1. Understanding hooks and their events30 min
    2. Configuring a hook with /hooks and settings.json35 min
    3. PreToolUse: validating or blocking an action before it runs45 min
    4. PostToolUse: formatting and checking code automatically40 min
    5. Lifecycle hooks: SessionStart, Stop and Notification40 min
    6. Securing and debugging your hooks35 min

  3. Module 3

    Plugins

    Bundle commands, subagents, skills, hooks and MCP servers into installable, shareable plugins.

    1. What is a Claude Code plugin?25 min
    2. Installing plugins from a marketplace with /plugin30 min
    3. Packaging commands, subagents, skills and hooks into a plugin50 min
    4. Creating and sharing a team marketplace40 min
    5. Governing plugin use across an organization30 min

  4. Module 4

    Advanced permissions and sandboxing

    Precisely limit what the agent can read, modify, execute and connect to.

    1. Applying the principle of least privilege to agents30 min
    2. Fine-grained permission rules for Bash, Read, Edit and WebFetch45 min
    3. Enforcing policy with managed settings35 min
    4. Isolating execution: sandboxing, containers and devcontainers50 min
    5. Controlling network and working-directory access40 min
    6. Workshop: hardening a project's configuration75 min

  5. Module 5

    Prompt injection and secret protection

    Recognize attacks delivered through untrusted content and protect sensitive information.

    1. Understanding direct and indirect prompt injection35 min
    2. Recognizing untrusted content: web, files and tool output35 min
    3. Defending against injection: isolation, validation and human confirmation45 min
    4. Protecting secrets: .env files, deny rules and environment variables40 min
    5. Detecting and revoking an exposed secret35 min
    6. Assessing the risks of third-party MCP servers35 min

  6. Module 6

    Code review and security

    Keep human control over AI-generated code and establish safe team practices.

    1. Systematically reviewing AI-generated code40 min
    2. Running a security review with /security-review35 min
    3. Integrating static analysis and security tests into CI45 min
    4. Tracing and auditing agent activity35 min
    5. Writing an AI usage policy for your team40 min
    6. Final project: a secure, well-equipped Claude Code environment160 min

Your way to learn

Three ways to complete a level

Personal guidance, the encouragement of a small group or self-paced learning: choose the format that fits your budget and availability.

Skills, hooks, permissions and security

Live · One-to-one

One instructor and one student on a video call, with personalized guidance.

MXN 5,000.00

per student · for one complete level

  • One-to-one sessions with your instructor.
  • Explanations adapted to your pace and goals.
  • Guided exercises, questions and feedback.
  • Progress through the selected level curriculum.

Times are agreed with the instructor. The schedule, number and duration of sessions are confirmed before payment.

Reserve with 500 MXN (10%) and pay the balance at the start. 50/50 option: 2,500 MXN at the start, including the deposit, and 2,500 MXN five weeks after the first scheduled class.

Request available times
Recommended format

Live · Small group

Learn together, ask questions and stay motivated by your peers at a more accessible price.

MXN 1,000.00

per person · for one complete level

  • 5 to 10 students in the same live class.
  • Discussion, peer support and group exercises.
  • Instructor demonstrations and questions and answers.
  • Several time slots may be proposed to match common availability.

Starts after at least 5 participants are confirmed; 10 maximum. The shared schedule is confirmed before payment.

Reserve with 100 MXN (10%) and pay 900 MXN at the start. The group needs 5 to 10 confirmed participants.

Join a group

Recorded videos · Self-paced

Study at your own pace through a curriculum organized by level, module and lesson, then put your learning into practice.

Price depends on the level

price for the selected level or full path

  • Recorded videos according to the level’s published content.
  • Detailed lessons, objectives and practical examples.
  • Exercises, submission instructions and quizzes.
  • Progress tracking and a level validation project.

No live appointment: organize your own study time. Content availability is shown in the curriculum.

View recorded learning

Live prices are in Mexican pesos (MXN) and cover one complete level. Recorded learning keeps its configured level or full-path price and the currency shown at checkout.

Live classes · Google MeetExplore the steps and course process