The Agent Skills Directory

[SAFE]: The skill is a technical documentation and workflow guide for auditing MCP servers. It promotes security-first development by checking for least-privilege access, origin validation, and input sanitization in the audited servers.- [COMMAND_EXECUTION]: The workflow includes a step to run repo-local integration tests and validation commands. This is the intended functional behavior for a developer auditing tool and is used to verify server compliance.- [INDIRECT_PROMPT_INJECTION]: As an auditing tool, this skill possesses an inherent attack surface for indirect prompt injection because it processes external server definitions and repository source code. Ingestion points: Server metadata, tool/prompt schemas, and source code files. Boundary markers: Not explicitly defined within the workflow instructions. Capability inventory: The skill uses file system access and shell command execution to perform audits. Sanitization: The skill instructions explicitly direct the agent to audit for sanitization in the target server, though it relies on the underlying LLM's inherent safety guardrails for its own input processing.

mcp-audit