signal-sourcer
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on executing shell commands via the
clayCLI tool to perform environment checks, manage workflows, and execute company searches. Examples includeclay whoamiandclay workflows creatementioned inSKILL.mdandreferences/detector-mechanics.md. - [INDIRECT_PROMPT_INJECTION]: The skill harvests data from external, untrusted sources such as Google News results and digest pages, which could theoretically contain instructions designed to influence the agent's behavior.
- Ingestion points: Article headlines, text, and metadata processed during the 'Harvest candidates' phase (Step 4 in
SKILL.md). - Boundary markers: Absent. There are no specific delimiters or instructions provided to the agent to prevent it from obeying instructions that might be embedded in the crawled news content.
- Capability inventory: The agent has the ability to execute shell commands (
clayCLI) and read local files (such as 'the book' of existing customers or 'answer sheets'). - Sanitization: The skill does not define explicit sanitization or validation logic for strings harvested from the web before they are used in subsequent search or resolution steps.
Audit Metadata