attio-hello-world
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where it ingests data from the Attio API to inform the creation of code and test fixtures.
- Ingestion points: External data is retrieved from the Attio REST API (e.g.,
https://api.attio.com/v2/objects) as described in the Instructions and Tool Discipline sections ofSKILL.md. - Boundary markers: The instructions do not specify the use of delimiters or explicit boundary markers for the agent when processing the received API data to ensure instructions embedded in CRM data are not executed.
- Capability inventory: The skill utilizes the
WriteandEdittools to modify the local repository by adding request wrappers and creating fixture tests based on API responses. - Sanitization: The instructions advise the agent to record only object slugs or counts that are safe to expose, which acts as a manual filtering step for external content.
Audit Metadata