rss-source-curator
Warn
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions guide the agent to execute shell commands using the Node.js runtime on local files, including
node scripts/doctor.mjsandnode scripts/bootstrap-runtime.mjs. This behavior enables the agent to execute code defined within the repository's script files for environment diagnostics and bootstrapping purposes. - [COMMAND_EXECUTION]: The agent is instructed to resolve its runtime command from environment variables like
DIGESTPILOT_RUNTIME_CMDor system paths, which could be leveraged to execute unexpected binaries if the environment is misconfigured or manipulated. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface as it processes external data from RSS feeds and JSON registry files (
feeds.json,source-health.json). Instructions or malicious content embedded within these data sources could attempt to influence the agent's curation logic or output reports, although the skill includes a safety boundary advising against automated feed deletion.
Audit Metadata