pdp-ecomm-template

Warn

Audited by Gen Agent Trust Hub on Jul 25, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill's 'Preamble' executes multiple local binaries/scripts: rmbc-update-check, rmbc-analytics, and rmbc-workspace. These tools manage skill updates and workspace configuration but introduce significant local execution surface.\n- [REMOTE_CODE_EXECUTION]: The skill implements an update mechanism that checks for new versions and directs the agent to follow 'Inline upgrade' instructions from potentially updated files in the rmbc-upgrade directory, allowing the behavior of the skill to change dynamically via external files.\n- [DATA_EXFILTRATION]: Usage analytics are logged via a background process (rmbc-analytics). Although an opt-in/opt-out mechanism is provided via an interactive prompt, this represents a network operation that sends telemetry data regarding skill and product usage.\n- [PROMPT_INJECTION]: The skill ingests untrusted content from the reference_page_url input parameter. This creates a surface for indirect prompt injection where instructions hidden in the external webpage could attempt to override the agent's behavior.\n
  • Ingestion points: reference_page_url input parameter in SKILL.md.\n
  • Boundary markers: The skill does not define delimiters or instructions to ignore embedded commands in the reference content.\n
  • Capability inventory: The skill has the ability to execute shell commands and modify local files in the ~/.rmbc-skills directory.\n
  • Sanitization: No sanitization or validation of the fetched webpage content is described.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 25, 2026, 12:14 PM
Security Audit — agent-trust-hub — pdp-ecomm-template