verify-form-validation
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to initialize the Reticle environment using
npx @reticlehq/server@latest initand fetches additional technical context from the vendor's official domain viacurl https://docs.reticle.sh/capabilities.mdandcurl https://docs.reticle.sh/llms.txt. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates interaction with external web content, which presents a potential surface for indirect prompt injection if the target application contains malicious instructions designed to influence the agent's behavior during the test.
- Ingestion points:
reticle_look(extracts page state and element references),reticle_act_and_wait(interpolates values into the browser session). - Boundary markers: No explicit delimiters or instructions to ignore embedded content are defined in the provided code snippets.
- Capability inventory:
reticle_act,reticle_act_and_wait, andreticle_run(tools for browser interaction and execution flow control). - Sanitization: No specific sanitization or filtering logic for ingested page data is mentioned in the instructions.
Audit Metadata