Audit local agent skills for supply-chain, prompt-injection, secret-leakage, dependency-install, webhook, destructive-shell, manifest-quality, duplicate-name, auxiliary-doc, and resource-discovery risks.
AI agents call audit_installed_skills to retrieve information from CodeAudit MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool audits/inspects local agent skills for various risk categories. The server is explicitly described as 'read-only' and the tool performs analysis/inspection rather than modifying or executing anything. However, severity is medium because it reads potentially sensitive data (secrets, agent skills, manifests) and its findings could inform further actions.
From the tool's definition Audit local agent skills for supply-chain, prompt-injection, secret-leakage, dependency-install, webhook, destructive-shell, manifest-quality, duplicate-name, auxiliary-doc, and resource-discovery risks.
Attacks that exploit this kind of access
Audit local agent skills for supply-chain, prompt-injection, secret-leakage, dependency-install, webhook, destructive-shell, manifest-quality, duplicate-name, auxiliary-doc, and resource-discovery risks. It is categorised as a Read tool in the CodeAudit MCP MCP Server, which means it retrieves data without modifying state.
Register the CodeAudit MCP server in PolicyLayer and add a rule for audit_installed_skills: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches CodeAudit MCP. Nothing to install.
audit_installed_skills is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the audit_installed_skills rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for audit_installed_skills. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
audit_installed_skills is provided by the CodeAudit MCP server (priyanshuchawda/codeaudit). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →