process-dedup-review
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a Node.js wrapper script (
run-process-dedup-review.mjs) to execute thetiangong-lcaCLI tool. This is the primary mechanism for performing the deduplication review and remote metadata enrichment. The tool is a vendor-owned resource and is used in a manner consistent with its intended purpose. - [INDIRECT_PROMPT_INJECTION]: The skill processes external JSON data containing process information, which could serve as a potential surface for indirect injection if malicious instructions were embedded in the metadata.
- Ingestion points: Data is ingested from a user-supplied JSON file provided via the
--inputargument as described inSKILL.md. - Boundary markers: None explicitly implemented in the skill instructions, though the use of a strict "Canonical Input Contract" for JSON data provides structural delimitation.
- Capability inventory: The skill invokes shell commands via the
tiangong-lcaCLI and performs file write operations to an output directory. - Sanitization: The skill relies on the external
tiangong-lcaCLI tool to validate and sanitize input data before processing.
Audit Metadata