keyword-gap
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from the DataForSEO API (keyword and competitor data) which is subsequently processed and clustered by the agent.
- Ingestion points: The skill uses
scripts/dfs_client.pyto fetch keyword lists and competitor metrics from the DataForSEO API. - Boundary markers: The instructions do not specify explicit delimiters or "ignore instructions" wrappers for the ingested keyword data.
- Capability inventory: The skill has the ability to execute local Python scripts (
dfs_client.py,project_memory.py,report_publish.py) and write output to the local file system in Markdown and PDF/HTML formats. - Sanitization: No explicit sanitization or filtering of API-returned text strings is defined in the instructions.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute several local Python scripts, passing user-provided domain names and competitor names as command-line arguments.
- Evidence: Found in the "Data pulls" and "Output" sections:
python scripts/dfs_client.py competitors --target "<user-domain>",python scripts/dfs_client.py intersection --target1 "<user-domain>" --target2 "<competitor-a>", andpython scripts/report_publish.py <report>.md. - Note: While placeholders are used, the interpolation of user-supplied domain strings into shell commands represents a standard surface for potential argument injection if not handled by the platform's execution environment.
Audit Metadata