Security category

Claude Code Security Mods

Claude Code security mods include prompt redaction, path guards, and tool checks. Compare their source evidence and interception scope.

A guardrail can miss things. This category is not a safety certification, and a mod runs with your user permissions.

Installation guide ↗

Claude Code Security Mods directory

7 mods
MOD-015Security
  • tool.call

secret-redactor

Masks credential patterns in tool results before model consumption.

by claude-code-templates

Permissions Facts

Source 8da17d67 · 2026-10-08

Can read files
Not declared
Can write files
Not declared
Can access the network
Not declared
Can read environment variables
Not declared
Can change prompts
Not declared
Can intercept tool calls
Yes

Capability evidence, not a security certification.

Not listed by claude plugin validate; not a guarantee.

Review pinned module ↗
32,476repository stars2026-10-08repository commit
MOD-005Security
  • tool.call

block-destructive-commands

Blocks Bash commands matching configured destructive patterns.

by claude-code-templates

Permissions Facts

Source 8da17d67 · 2026-10-08

Can read files
Not declared
Can write files
Not declared
Can access the network
Not declared
Can read environment variables
Not declared
Can change prompts
Not declared
Can intercept tool calls
Yes

Capability evidence, not a security certification.

Not listed by claude plugin validate; not a guarantee.

Review pinned module ↗
32,476repository stars2026-10-08repository commit
MOD-007Security
  • prompt.submit
  • turn.step

jev-guardrails

Screens prompts and responses using configurable Jev policies.

by claude-code-templates

Permissions Facts

Source 8da17d67 · 2026-10-08

Can read files
Not declared
Can write files
Not declared
Can access the network
Yes
Can read environment variables
Not declared
Can change prompts
Yes
Can intercept tool calls
Not declared

Capability evidence, not a security certification.

Not listed by claude plugin validate; not a guarantee.

Review pinned module ↗
32,476repository stars2026-10-08repository commit
MOD-009Security
  • session.start
  • tool.call

large-edit-confirmation

Requests confirmation before changing files over a size threshold.

by claude-code-templates

Permissions Facts

Source 8da17d67 · 2026-10-08

Can read files
Yes
Can write files
Not declared
Can access the network
Not declared
Can read environment variables
Not declared
Can change prompts
Not declared
Can intercept tool calls
Yes

Capability evidence, not a security certification.

Not listed by claude plugin validate; not a guarantee.

Review pinned module ↗
32,476repository stars2026-10-08repository commit
MOD-014Security
  • tool.call

protected-paths-guard

Blocks file-editing tools on configured sensitive paths.

by claude-code-templates

Permissions Facts

Source 8da17d67 · 2026-10-08

Can read files
Not declared
Can write files
Not declared
Can access the network
Not declared
Can read environment variables
Not declared
Can change prompts
Not declared
Can intercept tool calls
Yes

Capability evidence, not a security certification.

Not listed by claude plugin validate; not a guarantee.

Review pinned module ↗
32,476repository stars2026-10-08repository commit
MOD-028Security
  • session.start
  • command.run
  • tool.call
  • ui.render

guardrails

Provides configurable rules for risky commands and file access.

by Michael Goldenberg

Permissions Facts

Source 0954c978 · 2026-10-08

Can read files
Not declared
Can write files
Not declared
Can access the network
Not declared
Can read environment variables
Not declared
Can change prompts
Not declared
Can intercept tool calls
Yes

Capability evidence, not a security certification.

Not listed by claude plugin validate; not a guarantee.

Review pinned module ↗
5repository stars2026-10-06repository commit
MOD-043Security
  • session.start
  • tool.call
  • command.run
  • ui.render

launch-codes

Adds an authorization-code workflow for dangerous Bash commands.

by OneWave AI

Permissions Facts

Source e6da26ca · 2026-10-08

Can read files
Not declared
Can write files
Not declared
Can access the network
Not declared
Can read environment variables
Not declared
Can change prompts
Not declared
Can intercept tool calls
Yes

Capability evidence, not a security certification.

Not listed by claude plugin validate; not a guarantee.

Review pinned module ↗
7repository stars2026-10-03repository commit

Repository metrics captured 2026-10-08. Listings document source evidence; installation and runtime behavior have not been tested.

Frequently asked questions

What should I check before choosing a security mod?

A guardrail can miss things. This category is not a safety certification, and a mod runs with your user permissions.

Are these mods safe to install?

This directory documents source evidence, not safety guarantees. Mods run with your user permissions. Review the pinned source, dependencies, and permission facts before installing.

What do the stars and dates mean?

Stars are GitHub repository stars at the recorded snapshot. Repository commit is the date of the captured repository HEAD, not the last update to an individual mod.

Does this website install a mod for me?

No. Search stays in your browser, and copying commands does not run them. You review the source and decide what to install in your own Claude Code session.