launch-feedback-synthesizer
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data, including community comments, store reviews, and support tickets. This ingestion represents a potential attack surface for indirect prompt injection. The skill mitigates this by including explicit instructions to treat all feedback as data to be clustered rather than instructions to be followed, referring to an internal security policy.
- [COMMAND_EXECUTION]: The skill references the execution of local scripts for data acquisition and registry updates, such as
scripts/connectors/hn.py,scripts/connectors/appstore.py, andregistry-events.py. These scripts are used for specific, scoped tasks like fetching public thread data or proposing data entries to an internal registry. - [DATA_EXPOSURE]: The skill accesses and writes to local memory directories (e.g.,
memory/launch/,memory/events/). It operates under a restricted writing model where it can only propose changes to the central registry through a separate script (registry-events.py) rather than modifying the registry directly, adhering to the principle of least privilege.
Audit Metadata