skills/jeremylongshore/claude-code-plugins-plus-skills/apollo-reference-architecture/Gen Agent Trust Hub
apollo-reference-architecture
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface detected. The skill processes external data from the Apollo API in 'SKILL.md' (e.g., in 'enrichLead' and 'searchAndEnrich' methods) without boundary markers or explicit sanitization to prevent embedded instructions from influencing the agent.
- [COMMAND_EXECUTION]: The skill requests broad permissions for 'Bash(gh:)' and 'Bash(curl:)' in its frontmatter. While no malicious usage is present in the documentation, these tools provide significant system and network access capabilities.
- [DATA_EXFILTRATION]: The architecture handles sensitive Personal Identifiable Information (PII) such as contact emails and LinkedIn profiles. The combined availability of network tools like 'curl' establishes a potential path for data exfiltration that should be monitored.
Audit Metadata