Skip to content

Concepts

This page explains the model behind the CLI. Open the concepts overview

Hooks

Concept

envctl can manage a very small Git-hook safety layer. The point is not generic automation. The point is narrow local protection against committing or pushing envctl-managed secret material by mistake.

What hooks are

In the envctl model, managed hooks are local wrappers around:

envctl guard secrets

They exist to stop obvious secret-handling mistakes before commit and push.

Why they matter

Hooks add a lightweight local safety net:

  • before a commit is created
  • before a push leaves the machine

That makes them useful, especially for teams, but they remain a narrow protection layer.

What problem they solve

Managed hooks solve one focused problem:

prevent accidental local Git operations from carrying envctl-managed secret material forward unnoticed

They do not try to solve broader CI policy, arbitrary automation, or multi-tool hook orchestration.

What hooks are not

Managed hooks are not:

  • a general-purpose automation framework
  • a hook-merging system
  • a replacement for CI enforcement
  • a guarantee against --no-verify

That narrow scope is intentional. It avoids hidden behavior and fragile integrations.

Managed vs foreign hooks

envctl distinguishes:

  • managed hooks: created by envctl, marked as managed-by: envctl, and fully controlled by it
  • foreign hooks: created by something else or manually modified

If a hook is not managed by envctl, it is left alone by default.

How hooks fit in the system

Managed hooks sit beside the core model, not inside it:

  • the contract defines requirements
  • the vault stores local values
  • resolution and projection govern runtime truth
  • hooks add a local Git safety layer around that model

Using hooks

See the operational workflow for install, status, repair, and removal.

Open hooks guide

hooks reference

See the exact command surface for the hook management commands.

Open hooks reference

Security reference

Connect hook protection back to the broader security model and its limits.

Open security reference