fusion-dependency-review
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content, including pull request comments, review threads, and upstream changelogs. This represents a potential surface for indirect prompt injection where malicious instructions could be embedded in the data being reviewed. However, the skill implements strong mitigations by requiring the agent to cite evidence for all claims, providing structured templates for analysis, and strictly mandating explicit maintainer confirmation before any merge, approval, or branch mutation occurs.
- [COMMAND_EXECUTION]: The skill is mutation-capable and performs git operations such as rebase, sync, and merge via the GitHub MCP. These operations are guarded by strict safety constraints: the agent must never merge or approve without explicit user confirmation, must prefer rebase over merging base branches into bot-authored PRs, and must ask for confirmation before any force-push or merge action.
Audit Metadata