managing-snapshot-tests
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes UI component outputs and serialized data which can originate from external sources. If these inputs contain malicious instructions, they could influence the agent's behavior during the snapshot analysis or update process.
- Ingestion points: Component rendering output (HTML, JSON, YAML, TOML) described in the SKILL.md instructions.
- Boundary markers: None specified in the instructions for delimiting or flagging the processed output.
- Capability inventory: The skill utilizes the Write, Edit, and Bash tools to manipulate snapshot files and run tests.
- Sanitization: No specific sanitization or filtering of the component output is mentioned before processing.
Audit Metadata