agent-update-manager
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from a remote repository, such as changelogs and version metadata.
- Ingestion points: External data is retrieved in Step 4 and Step 7 of the
SKILL.mdworkflow. - Boundary markers: Human-in-the-loop approval is mandatory (Step 8 and 9) before any updates are executed.
- Capability inventory: The skill utilizes
capabilities/file-operations,capabilities/web-research, andcapabilities/project-knowledge-graphto assess and apply updates. - Sanitization: A dedicated risk assessment step (Step 7) is performed to detect breaking changes or security policy alterations.
- [EXTERNAL_DOWNLOADS]: The skill fetches update data from the author's official repository at
github.com/lensetek/Digital-Marketing-Agent_Skills. This is an expected and functional use of a vendor-owned resource.
Audit Metadata