lca-publish-executor
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes the
tiangong-lcaCLI tool using a Node.js wrapper script (scripts/run-lca-publish-executor.mjs). This is the intended functionality for orchestrating publication tasks through official vendor tooling. - [INDIRECT_PROMPT_INJECTION]: The skill processes artifact bundles and dataset payloads generated by other upstream skills, which presents a surface for processing untrusted data.
- Ingestion points: External data is ingested via file paths in
inputs.bundle_pathsor direct JSON payloads in theinputsobject of the request manifest. - Boundary markers: The skill implements strict JSON schema validation via
assets/request.schema.jsonand instructs the agent to inspectverification-report.jsonfor blockers before proceeding. - Capability inventory: The skill facilitates network write operations to the vendor's platform through the
tiangong-lca publish runcommand. - Sanitization: Input is validated against a formal schema, and the underlying CLI tool performs independent verification of all artifact payloads.
Audit Metadata