selfhosted-stack-analyzer
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on standard administrative tools such as
kubectl,aws CLI,jq, andhelmto perform Phase 1 data collection. These operations are limited to read-only diagnostic commands (get,list,describe) to ensure the security and integrity of the audited environment. - [EXTERNAL_DOWNLOADS]: The
generate-html-report.pyscript generates an HTML report that includes a reference to load the Mermaid.js library from a public CDN (https://cdn.jsdelivr.net/npm/mermaid@10/dist/mermaid.min.js). This is used to render architectural topology diagrams locally in the user's browser. - [INDIRECT_PROMPT_INJECTION]: The skill ingests infrastructure metadata such as resource names and EC2 instance tags. While this represents a potential surface for indirect prompt injection, the skill incorporates strong mitigations through its 'Evidence-Driven' hard rules, which require the agent to categorize all findings into specific evidence tiers (CONFIRMED, DECLARED, INFERRED) and strictly prohibit the fabrication of system roles or relationships without host-level proof.
Audit Metadata