prospect-refresh
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources (Apollo People API and Organization Enrichment) and interpolates this content into Gmail drafts and reports. The absence of explicit sanitization or boundary markers creates a surface for indirect prompt injection, where malicious content in an external prospect's record could attempt to influence the agent's behavior during the drafting or reporting process.\n
- Ingestion points: External prospect data and company enrichment details are retrieved via the
apollo_mixed_people_api_searchandapollo_organizations_enrichtools in Stages 2 and 5 of the workflow inSKILL.md.\n - Boundary markers: The skill uses simple bracketed placeholders (e.g.,
[FIRST_NAME],[COMPANY_NAME]) for template substitution without additional instructions to the model to ignore any embedded commands or special characters within the data.\n - Capability inventory: The skill possesses the capability to create Gmail drafts (
gmail_create_draft) and write to the local file system (generating an HTML report and writing a JSON analytics file to~/.claude/skill-analytics/).\n - Sanitization: No specific sanitization, escaping, or validation logic is defined in the skill instructions to process or filter the data retrieved from external APIs before use.
Audit Metadata