The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions in SKILL.md require the agent to download the @nomadamas/k-skill package from the NPM registry to obtain the full instruction set and manage skill files.
[REMOTE_CODE_EXECUTION]: The skill instructs the agent to run npx -y @nomadamas/k-skill@0, which results in the execution of code from an external, untrusted source at runtime.
[COMMAND_EXECUTION]: The agent is prompted to execute several shell commands, including npx for skill management and a local Python script (scripts/animal_pharmacy_mcp.py) to interface with the pharmacy database.
[DATA_EXFILTRATION]: The skill sends user-provided search keywords and geographic location information (city, district) to the external domain hkmedi.co.kr through HTTP POST requests initiated by the Python helper script.
[INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted data from a remote API which could potentially contain malicious instructions.
Ingestion points: The scripts/animal_pharmacy_mcp.py script retrieves and parses JSON/SSE data from https://hkmedi.co.kr/pharmacy-mcp.
Boundary markers: The instructions do not define delimiters or specific warnings to ignore instructions that might be embedded within the API's pharmacy or product data.
Capability inventory: The skill possesses shell execution capabilities (via npx and python scripts) and network access.
Sanitization: While the script validates that the response is valid JSON, it does not perform sanitization or filtering of the content for natural language injection patterns.

animal-pharmacy-search