skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/clash-resolution-analyzer/Gen Agent Trust Hub
clash-resolution-analyzer
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external clash detection data provided by the user.
- Ingestion points: According to
instructions.md, the agent accepts project data in formats such as CSV, Excel, and JSON, or via direct input. This data is processed by theimport_clashesmethod inSKILL.md. - Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from following malicious commands hidden within the processed data (e.g., instructions embedded in the 'resolution_notes' or 'name' fields of a clash report).
- Capability inventory: The skill metadata in
claw.jsonrequestsfilesystempermissions. The Python implementation inSKILL.mddefines methods for data manipulation and report generation which are intended to be executed by the agent. - Sanitization: The
import_clashesmethod inSKILL.mdperforms basic type checks viadata.get()but does not sanitize the string content of the data against prompt injection attacks.
Audit Metadata