dark-funnel
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it aggregates data from untrusted external channels.
- Ingestion points: Data enters the context from Reddit discussions, community forums, podcasts, and survey data mentioned in the 'Aggregate dark funnel signals' process step.
- Boundary markers: The instructions lack delimiters or safety warnings to ignore instructions that might be embedded in the aggregated marketing data.
- Capability inventory: The skill utilizes market-intelligence and analytics agents to process data, though it does not explicitly call subprocesses or execute arbitrary code.
- Sanitization: There is no mention of sanitization or validation of the external content before it is processed by the agent.
- [DATA_EXFILTRATION]: The skill performs read operations on the local file system to load marketing context and SOPs.
- Evidence: The skill reads brand-specific profiles and guidelines from
~/.claude-marketing/brands/and~/.claude-marketing/sops/.
Audit Metadata