blog-flow
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
/blog flow synccommand downloads updated SEO prompt templates and reference documentation from the author's GitHub repository (github.com/AgriciDaniel/flow). This allows the skill to incorporate the latest FLOW framework updates.- [COMMAND_EXECUTION]: The synchronization process is managed by executing a local Python script (scripts/sync_flow.py).- [PROMPT_INJECTION]: The skill is designed to ingest and analyze content from external URLs supplied by the user. This exposes the agent to indirect prompt injection risks if the targeted web content contains malicious instructions intended to manipulate the analysis output. - Ingestion points: URL parameters in
/blog flow find,/blog flow optimize, and/blog flow wincommands. - Boundary markers: None described in the orchestration logic to delimit untrusted web content.
- Capability inventory: File system access for reading references and executing the sync script.
- Sanitization: No explicit sanitization or validation of the retrieved URL content is mentioned.
Audit Metadata