launch-tier-planner
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill implements proactive security measures by instructing the agent to treat all user-provided plans, partner documents, and exports as untrusted input, specifically warning against following instructions embedded within those data sources.
- [COMMAND_EXECUTION]: The skill references the use of local scripts (
scripts/connectors/hn.pyandscripts/connectors/gdelt.py) to fetch public telemetry data from Hacker News and GDELT. These are used for legitimate market research and are part of the skill's defined infrastructure. - [DATA_EXFILTRATION]: File access is restricted to the local environment and project-specific
memory/directories. While the skill processes analytics exports, it does not demonstrate any patterns of sending sensitive data to unauthorized external domains.
Audit Metadata