seo-traffic-drop
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources that could contain adversarial instructions targeting the agent.
- Ingestion points: The skill is designed to ingest and process data from Search Console CSV exports (
Queries.csv,Pages.csv,Dates.csv) and local markdown notes within an Obsidian vault. - Boundary markers: There are no explicit boundary markers (e.g., XML tags or delimiters) or instructions provided to the agent to treat the contents of these files strictly as data, increasing the risk that embedded instructions could be executed by the agent upon reading script outputs.
- Capability inventory: The skill allows for the execution of local Python scripts and shell commands (
curl) based on the diagnosis results, which could be exploited if an injection is successful. - Sanitization: While the Python scripts include basic error handling for numeric parsing, they do not sanitize or escape the string content (like query terms or page titles) retrieved from the CSV files before outputting them for the agent to read.
- [DYNAMIC_EXECUTION]: One of the provided scripts uses dynamic module loading to manage internal dependencies.
- Evidence: The file
scripts/content_decay.pyuses theimportlib.utillibrary to dynamically loadscripts/gsc_diff.py. While this is implemented to allow code reuse between local scripts, dynamic loading of executable code is a pattern that can be misused if paths are manipulated. In this case, the path is derived from the script's own location.
Audit Metadata