megatron-change-analyzer
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external commit metadata which could be used as a vector for indirect prompt injection.
- Ingestion points: The skill reads commit titles and changed file lists via the
--change-setparameter inscripts/build_feature_events.pyand subsequently processes them via instructions inSKILL.md. - Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat the ingested data as untrusted, which may lead the agent to interpret embedded instructions within commit messages as valid commands.
- Capability inventory: The skill is limited to local file processing and JSON transformation. No high-risk capabilities like network access, arbitrary shell execution, or dynamic code evaluation were identified in the provided scripts.
- Sanitization: The script
scripts/build_feature_events.pyfilters certain "noise" commits based on titles but lacks specific sanitization logic to detect and neutralize prompt injection attempts within the commit data.
Audit Metadata