social-creative-builder
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from user-provided articles, exports, and scraped web pages, creating a risk for indirect prompt injection attacks.
- Ingestion points: The skill ingests user-provided ideas, source articles, analytics exports, and scraped web pages as noted in the Instructions and Data Sources sections.
- Boundary markers: The skill includes explicit security instructions directing the agent to treat these sources as untrusted input and prevents them from approving claims or waiving disclosures.
- Capability inventory: The skill possesses the capability to read internal project files, write to local memory, and execute localized scripts for network connectivity and event registration.
- Sanitization: The skill relies on model-level instructions for sanitization; it does not describe programmatic filtering or escaping of the ingested content.
- [COMMAND_EXECUTION]: The skill relies on local script execution to interact with external search services and project registries.
- Evidence: SKILL.md references the execution of localized scripts including scripts/connectors/tavily.py, scripts/connectors/bluesky.py, and registry-events.py.
Audit Metadata