content-writer
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a local Python script
indexpush.pyto notify search engines (Bing, DuckDuckGo, Baidu) when content is updated. This operation is restricted to final, live URLs and utilizes standard environment variables ($INDEXNOW_KEY,$BAIDU_PUSH_TOKEN) for authentication, which is an expected behavior for a marketing-oriented skill. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, such as URLs and third-party CSV exports, which creates a potential surface for indirect prompt injection. However, the skill includes explicit, high-priority instructions for the agent to treat this data as untrusted and to never follow instructions embedded within fetched content, effectively mitigating the risk.
- [DATA_EXPOSURE_&_EXFILTRATION]: The skill writes content drafts and refresh reports to local memory paths (
memory/content/content-writer/). While it utilizes API keys via environment variables for indexing, these are passed to a local connector script rather than being hardcoded or sent to arbitrary external domains.
Audit Metadata