value-first-cold-email
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
claycommand-line utility to perform workspace identification (clay whoami), credit monitoring (clay credits), and data lookup routines (clay routines runs start). These commands are essential for the skill's primary function and are subject to explicit user-defined credit limits. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, which introduces a potential surface for indirect prompt injection.
- Ingestion points: The agent reads from prospect lists (Step 1) and external signal payloads retrieved via Clay API calls (Step 5), including news text and job descriptions.
- Boundary markers: There are no explicit structural markers used to isolate ingested data from the prompt instructions.
- Capability inventory: The skill has the ability to write to the conversation, write to local files, and execute specific CLI tools, but is restricted from sending emails or performing automated enrollments.
- Sanitization: The skill applies a 'Copy checks' methodology (referenced in
references/copy-checks.md) to the generated output to filter out common AI-generated filler words and spam patterns, though it does not explicitly sanitize inputs for adversarial instructions.
Audit Metadata