website-ai-agent-readiness
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates an indirect prompt injection attack surface by instructing the agent to ingest and process external website files.\n- Ingestion points: Step 2 of the workflow in SKILL.md identifies several external data sources including robots.txt, sitemap.xml, llms.txt, pricing.md, and ai-index.json.\n- Boundary markers: The skill lacks explicit instructions for the agent to treat external content as untrusted or to use secure delimiters when processing these files.\n- Capability inventory: The agent is empowered to modify project files and execute shell commands for build and smoke testing (SKILL.md, Step 6).\n- Sanitization: No validation or sanitization protocols are defined for the incoming external data.\n- [COMMAND_EXECUTION]: The skill directs the agent to execute shell commands to run regression and build tests.\n- Evidence: Step 6 of the workflow in SKILL.md instructs the agent to "Run the narrow SEO test suite first, then the broader build or smoke suite." This involves the execution of local shell commands which could be leveraged if the agent is influenced by malicious external content.\n- [NO_CODE]: The skill consists entirely of instructional Markdown and YAML configuration files and does not include any executable scripts, binaries, or logic-bearing code.
Audit Metadata