zipcode-search
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a Python script that invokes the
curlutility via thesubprocessmodule. This is a standard method for interacting with web services within the agent environment and is used here to perform address queries. - [EXTERNAL_DOWNLOADS]: The skill performs HTTP GET requests to the official South Korean postal service domain (
parcel.epost.go.kr) to retrieve postal code information. This connection is used strictly for the skill's documented purpose and targets a well-known public service endpoint. - [PROMPT_INJECTION]: The skill retrieves and processes external HTML data, which represents a potential surface for indirect prompt injection.
- Ingestion points: Data is fetched into the
pagevariable from the curl output inSKILL.md. - Boundary markers: No explicit instructions are used to separate the external content from the agent's instructions, though the data is filtered via regex.
- Capability inventory: The skill uses
subprocess.runto callcurland write results to stdout. - Sanitization: The skill uses
html.unescapebefore displaying results, though the raw HTML is primarily processed via structured regular expressions.
Audit Metadata