monitor-buying-signals
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
clayCLI commands to interact with the Clay platform, including environment verification (clay whoami), workspace checks (clay workflows actions list), and starting news routines (clay routines runs start). - [INDIRECT_PROMPT_INJECTION]: The skill ingests external news data (titles and snippets) which are then processed by an LLM for classification. This introduces a surface for indirect prompt injection where malicious content in news articles could attempt to influence the agent's behavior during the signal analysis phase.
- Ingestion points: External news events are pulled from Google News and managed news routines in Step 3 of
SKILL.md. - Boundary markers: No explicit delimiters or boundary markers are defined for the event content when passed to the LLM.
- Capability inventory: The agent has the ability to execute shell commands via the
clayCLI. - Sanitization: The skill employs deterministic keyword-based classification as a primary filter before using an LLM pass for ambiguous cases.
Audit Metadata