launch-day-conductor
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data, including community threads, dashboard screenshots, and metrics exports. It mitigates risk by including explicit instructions in the 'Instructions' section to treat these inputs as untrusted and to specifically ignore any embedded instructions within telemetry or comments.
- Ingestion points: Metric exports, dashboard screenshots, and community threads (specified in the 'Instructions' section).
- Boundary markers: Present. The skill instructs the agent to treat inputs per 'SECURITY.md' and ignore embedded instructions.
- Capability inventory: Writing runbook and incident logs to 'memory/launch/launch-day-conductor/' and proposing registry updates via 'registry-events.py'.
- Sanitization: Explicit instruction to disregard commands found within processed telemetry.
- [DATA_EXPOSURE]: The skill manages launch-related data within the 'memory/' directory. It utilizes a structured proposal system via 'registry-events.py' to update 'memory/events/launches.ndjson', maintaining a clear separation between proposed updates and canonical records. No hardcoded credentials or unauthorized data access patterns were detected.
- [EXTERNAL_DOWNLOADS]: The skill references several connector scripts located in 'scripts/connectors/' (e.g., 'hn.py', 'producthunt.py', 'appstore.py'). These are local scripts within the project's repository used for data fetching and do not constitute unauthorized remote code execution or downloads.
- [COMMAND_EXECUTION]: The skill interacts with local system components through predefined scripts like 'registry-events.py' and various connectors. These operations are limited to the skill's intended administrative and monitoring functions within the RAMP framework.
Audit Metadata