nonprofit-fundraising-configure
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill is designed to assign permission sets and permission set groups org-wide to all human users based on their profile. This mass modification of security settings is a high-privilege administrative action.
- [COMMAND_EXECUTION]: The skill employs the Salesforce CLI (
sf project deploy start) as a fallback path for deploying metadata fragments (Profile record type visibilities) if the headless-360 metadata operations fail or are unavailable. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by making configuration decisions based on untrusted data from the org.
- Ingestion points: Organization data including User lists, Profile names, and existing Record Types retrieved via
mcp__headless-360__dispatch_readonly(SKILL.md, references/mcp-invocation.md). - Boundary markers: The skill does not explicitly define delimiters for tool-returned data used in subsequent logic.
- Capability inventory: Extensive capabilities including permission set assignment, record creation, org-wide feature enablement, and CLI-based metadata deployment (SKILL.md).
- Sanitization: No specific sanitization or validation of retrieved string values (like Profile names) is documented before they are used to drive assignment logic.
- [SAFE]: Irreversible actions, such as enabling Person Accounts or Fundraising, are protected by explicit mandatory confirmation prompts (
AskUserQuestion) as defined in Phase 2.
Audit Metadata