inbound-lead-qualifier-skill
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from lead form fills and demo requests to generate scores and outreach templates. This data is external and potentially adversarial, which could lead to manipulation of the agent's logic.
- Ingestion points: Lead data, form fields, and engagement history ingested via the 'Trigger' and 'workflow' sections.
- Boundary markers: No specific boundary markers or delimiters are provided to separate user-provided lead data from the agent's instructions.
- Capability inventory: The skill possesses the capability to write files to the local file system.
- Sanitization: The instructions do not define any sanitization, filtering, or validation steps for the external lead data before processing.
- [COMMAND_EXECUTION]: The 'Emit Outcome Sidecar' section directs the agent to perform a file system write operation to a hidden path:
~/.claude/skill-analytics/last-outcome-inbound-lead-qualifier.json. While intended for analytics, writing to hidden directories in the home folder is a sensitive operation that lacks user-facing transparency.
Audit Metadata