openapi-endpoint-filtering
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a legitimate utility script for local file processing using standard Python libraries (json, pathlib). It does not perform network operations, exfiltrate data, or contain any malicious obfuscation. All patterns detected are consistent with the stated goal of optimizing token usage through API specification pruning.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external API specifications, which constitutes untrusted data. While an attacker could potentially embed malicious instructions in a specification's description or metadata fields, this is a general risk associated with processing any external documentation. The 'slicing' logic provided does not introduce new vulnerabilities. Mandatory evidence chain:
- Ingestion points: The provided Python script reads a user-specified OpenAPI file path using
spec_path.read_text()inSKILL.md. - Boundary markers: Absent; the script extracts and returns raw JSON segments.
- Capability inventory: The script uses standard file reading and JSON parsing; it does not include subprocess execution or network requests.
- Sanitization: Absent; the script extracts fields as-is, which is expected for a schema processing utility.
Audit Metadata