Claude Tools · Discovery + safety

Agents

Delegable roles with scoped tool grants. The card shows the role and the tools it can use — over-broad grants get flagged.


Audited listings

4 entries · pipeline-verified

CLEARED_STATIC

agent-critterdocs

jedeiah/agent-critter

一只桌宠小精灵,实时展示 Claude Code 的工作状态。支持 Petdex 社区 2700+ 精灵,透明窗口、可拖拽、可缩放。

Audited 2026-08-08View audit →

CLEARED_STATIC

ml-odyssey-skillssecurity

mvillmow/projectodyssey

Skills migrated to ProjectMnemosyne for cross-project reuse

Audited 2026-08-09View audit →

CLEARED_STATIC

digital-innovation-agentssecurity

pssah4/digital-innovation-agents

AI-augmented Innovation & Development Workflow: V-Model phases from business analysis to security audit

Audited 2026-08-09View audit →

CLEARED_STATIC

quarkus-agentdocs

quarkusio/quarkus-agent-mcp

MCP server for AI coding agents to create, manage, and interact with Quarkus applications. Provides tools for project scaffolding, dev mode lifecycle, extension

Audited 2026-08-09View audit →

Five illustrative examples — content shape only, not audited listings.

All examplesread-onlytestingsecurity

5 examples · starter content

Example 01

Code Reviewerread-only

Reviews a diff for correctness, security, maintainability, and project conventions.

Editable sampleView full page →

Example 02

Test Engineertesting

Designs focused tests for changed behavior and reports gaps without editing production code.

Editable sampleView full page →

Example 03

Security Auditorsecurity

Inspects code and configuration for injection, secrets, unsafe permissions, and risky defaults.

Editable sampleView full page →

Example 04

Documentation Editordocs

Improves technical documentation with scoped read and edit access.

Editable sampleView full page →

Example 05

Architecture Explorerexploration

Maps entry points, dependencies, and data flow using read-only repository tools.

Editable sampleView full page →