clickhouse-architect
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates the execution of local shell commands for database diagnostics and auditing.\n
- Evidence: The
SKILL.mdfile provides an instruction to execute the audit script usingclickhouse-client --multiquery < scripts/schema-audit.sql. This is standard diagnostic behavior for the skill's stated purpose.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from ClickHouse system tables, which represents an ingestion surface for potential indirect instructions if the database environment is compromised.\n - Ingestion points: Output from
scripts/schema-audit.sql, which queries metadata tables likesystem.query_logandsystem.processes.\n - Boundary markers: None. The output is processed directly without markers or instructions to ignore embedded commands.\n
- Capability inventory: The skill is configured with access to
Bash,Read,Grep, andSkilltools as defined inSKILL.md.\n - Sanitization: No evidence of sanitization or validation of the database output before it is processed by the agent.
Audit Metadata