ecc-claude-marketing-skill-analytics-tracking
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill implements an indirect prompt injection surface by design, as it instructs the agent to fetch and follow guidance from external files.
- Ingestion points: The workflow in
SKILL.mddirects the agent to read a 'reference snapshot' from a path derived fromreferences/upstream-path.txt(upstream-import/claude-skills/references/marketing-skill/analytics-tracking). - Boundary markers: There are no markers or instructions defined to prevent the agent from obeying malicious instructions that might be embedded within the upstream reference data.
- Capability inventory: The agent is instructed to 'Translate to OpenClaw tool-backed steps' and 'Execute incrementally', indicating it will use its available tools to perform actions based on the external content.
- Sanitization: The skill lacks any sanitization or validation logic to filter the content extracted from the upstream files before it influences the agent's execution path.
Audit Metadata