seo-task-generator
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external findings and reports that could contain malicious instructions designed to influence the agent's behavior.
- Ingestion points: Processes audit reports, analysis outputs, and files matching the pattern
SEO-AUDIT-*.md. - Capability inventory: The agent can write files to the filesystem (
TASKS-<domain>-<date>.md) and execute multiple local Python scripts. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands within the ingested audit data are defined.
- Sanitization: There are no instructions for validating or sanitizing the content of the external findings before processing.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local scripts using parameters derived from potentially untrusted inputs.
- Evidence: The skill calls
python scripts/recommend_store.py list --domain <domain>,python scripts/dfs_client.py onpage --url <sample>, andpython scripts/report_publish.py <report>.md. - Risk: Maliciously crafted domains, URLs, or file paths containing shell metacharacters could lead to command injection if not properly escaped before execution.
Audit Metadata