Security audit
Security checks for vulnerabilities and agentic risk
The skill provides travel searches through a disclosed third-party service, with no evidence of unrelated access or harmful actions.
Review and preferably pin the external package version, and protect your API key. Searches send supplied terms and locations to TripAdvisor; use a city or approximate location when sufficient. Enable the optional browser bridge only if you intend to permit browser-session use. This review covers the skill instructions, not a complete audit of the external package.
Skill accesses MCP server configuration files (mcp.json). MCP configs contain server URLs, authentication tokens, and tool definitions — reading them allows the skill to discover and potentially abuse other tool integrations.
Add to .mcp.json in your project or ~/.claude/mcp.json:
{
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
`ta_search_locations`, `ta_search_nearby`, `ta_get_locations` and
`ta_get_location_reviews` take `view: "compact" | "full"`, and **`compact` is
the default** — the slim rung is what you get without asking for it. Pass
`view: "full"` for TripAdvisor's whole records.
The compact rung is not the same operation on all four:
The workflow instructs the agent to run location or nearby searches, whose inputs can include user-provided coordinates or other location information. Although the file identifies the Terra API and describes operations as read-only, it does not warn that these inputs are transmitted to a third-party service; read-only access does not eliminate this privacy impact.
No suspicious patterns detected.