create-faceless-channel
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources such as APIs, feeds, and web pages to generate video scripts. This creates a vulnerability surface for indirect prompt injection, where an attacker-controlled source could embed instructions to manipulate the agent's output during content generation.
- Ingestion points: External data is ingested via the
source.entrypointURL defined in the channel configuration and processed during the 'Discover' and 'Capture' stages (seeSKILL.mdandreferences/configuration.md). - Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when passing retrieved content to the script generation phase.
- Capability inventory: The agent uses headless browser tools for page capture and language model capabilities to write video scripts based on the ingested facts.
- Sanitization: There are no documented sanitization or filtering steps for the external content before it is interpolated into generation prompts.
- [SAFE]: The skill follows security best practices for handling sensitive information. It explicitly instructs the agent and user to keep secrets in root-owned environment files, use scoped API keys, and never commit credentials to version control. It also enforces human gates for account creation and recovery, ensuring the agent does not automate high-risk identity tasks.
Audit Metadata