Skill Claude CodeCodex
A classification guide that assigns an assistant capability to the appropriate government business area using its purpose, functions, and evidence. It can assign more than one area when separate uses are supported.
Skill Claude CodeCodex
A safety-rating guide for coding-agent skills and tool connections. It classifies evidence about data access, permissions, credentials, external changes, and prompt attacks using fixed rules.
Skill Claude CodeCodex
A routing guide that turns a user's task into a focused search for relevant skills or tool connections. It creates stable search information while keeping the task's changing details separate.
Skill Claude CodeCodex
A compliance control for government work that checks roles, data labels, approvals, and legal duties before information is sent outside or exported. It only gives decisions and does not run workflows or approvals.
Skill Claude CodeCodex
A skill for drafting and checking official Chinese government documents such as notices, requests, and formal letters.
Skill Claude CodeCodex
A security service that decides whether an AI agent may call a tool. It uses user roles, task details, signed permissions, data labels, approvals, and audit records; risky requests can be stopped for human review.
Skill Claude CodeCodex
A meeting-notes guide that separates decisions, opinions, and unresolved questions, then extracts owners, deadlines, deliverables, and dependencies. Missing information is marked for confirmation.
Skill Claude CodeCodex
A controlled document tool for reading selected Word, Excel, PowerPoint, PDF, and text files and creating DOCX, XLSX, PPTX, PDF, or TXT files. It preserves document structure and checks generated files before delivery.
Skill Claude CodeCodex
A policy-reading assistant that finds policy clauses in user-provided material and separates quoted text, summaries, and conclusions. It analyzes who is affected, conditions, procedures, deadlines, responsibilities, and changes.
Skill Claude CodeCodex
A security checker for detecting prompt injection, jailbreak attempts, poisoned knowledge, and attempts to obtain sensitive information in enterprise AI-agent inputs.
Skill Claude CodeCodex
A guide for helping people complete government services by organising steps, required documents, conditional documents, and contact advice. The exact requirements can vary by location and time.
Skill Claude CodeCodex
A read-only classifier that decides whether a repository is a standard agent skill, an MCP server, both, or an ordinary repository. MCP is a protocol for connecting AI agents to tools, resources, and prompts.
Skill Claude CodeCodex
A policy skill for identifying sensitive information before sending data outside a system or exporting it. It covers credentials, personal information, internal labels, and data classifications.
Skill Claude CodeCodex
A security scanner for agent skills that examines packages without importing or running their code.
Skill Claude CodeCodex
An audit tool for recording versioned agent events, linking them with hashes and HMAC signatures, and protecting stored data through redaction or summaries. HMAC is a way to detect changes using a secret key.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: