seo-flow
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
/seo flow synccommand executes a local Python script (scripts/sync_flow.py) to download updated prompt files from the author's repository atgithub.com/AgriciDaniel/flow. This is a standard vendor-controlled update mechanism. - [COMMAND_EXECUTION]: The skill invokes shell commands including
python, and GitHub CLI tools (gh api rate_limit,gh auth login) to manage prompt synchronization and handle network-related errors during updates. - [PROMPT_INJECTION]: The skill processes external content from user-provided URLs and topics, which represents an indirect prompt injection surface. Maliciously crafted content on a target webpage could attempt to influence the agent's analysis.
- Ingestion points: User-supplied URL or topic arguments provided to the
/seo flow [stage]commands. - Boundary markers: Prompt templates in the
references/prompts/directory do not utilize specific delimiters (like XML tags) to isolate external content from the instruction set. - Capability inventory: The skill has capabilities for local file caching and command execution for maintenance, but these are not directly exposed to the content of analyzed URLs.
- Sanitization: There is no explicit sanitization or filtering of the text content ingested from user-provided URLs before it is processed by the analysis prompts.
Audit Metadata