marketing-project-mgmt
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and prioritize instructions from an external file named 'brand-context.md' that may be present in the user's project workspace.
- Ingestion points: In SKILL.md, the agent is instructed at Step 0 to search for and read 'brand-context.md' from the root, './.claude/brand-context.md', or './docs/brand-context.md' before any other tasks.
- Boundary markers: The skill explicitly specifies that the "Rules for agents reading this file" section within that external document "overrides an agent's own defaults," which creates a direct vector for untrusted content to hijack the agent's behavior.
- Capability inventory: The skill and its sub-agents (Campaign Coordinator, Resource Allocator, Scrum Master, and Client Success Manager) have significant capabilities to generate project plans, SOWs, and strategic reports based on this ingested context.
- Sanitization: The instructions lack any specific sanitization or validation protocols to prevent malicious instructions embedded in the 'brand-context.md' file from being executed by the AI.
Audit Metadata