The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill configuration and instructions promote the use of npx -y mcp-remote@latest to download and execute code from the npm registry at runtime. This bridge facilitates connection to the remote endpoint https://api.moodtrip.ai/api/mcp-http via a local subprocess.
[PRIVILEGE_ESCALATION]: The SKILL.md file contains instructions for the agent to autonomously "patch config so the MCP server is added automatically" if configuration tools are present. This allows the skill to modify the agent's environment and grant itself persistent command execution capabilities without explicit user confirmation of the configuration change.
[EXTERNAL_DOWNLOADS]: The skill requires the download of the mcp-remote package from the npm registry and relies on hotel inventory and pricing data from third-party domains api.moodtrip.ai and liteapi.travel.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from external APIs, such as guest reviews and AI-generated responses to hotel-specific questions, which could be manipulated to include malicious instructions targeting the agent.
Ingestion points: Data returned by getHotelReviews, ask_about_hotel, and search tools from api.moodtrip.ai described in SKILL.md and README.md.
Boundary markers: Absent; the instructions do not provide delimiters or warnings to ignore instructions embedded in the API data.
Capability inventory: The skill possesses network access capabilities and instructions for modifying platform configuration files to enable command execution.
Sanitization: Absent; no sanitization or validation of the external API content is described before it is processed by the agent.

moodtrip-hotel-search