deal-outreach
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes
findandgrepcommands to scan the local directory for markdown files containing specific business keywords (e.g., 'ICP', 'persona', 'value prop') to personalize the generated outreach sequences. - [PROMPT_INJECTION]: The skill demonstrates a surface for indirect prompt injection by ingesting content from local markdown files to calibrate outreach personas.
- Ingestion points: Content is read from local
.mdfiles identified via the context-gathering search commands (SKILL.md). - Boundary markers: The instructions do not define specific markers to isolate ingested project data from agent instructions.
- Capability inventory: The skill has access to
Bash,Read,Glob, andGrep(SKILL.md). - Sanitization: No validation or sanitization is performed on the content read from external files before it is processed by the agent.
Audit Metadata