improve-animations
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill contains a defensive instruction (Hard Rule 4) that correctly identifies 'ignore previous instructions' as a potential attack pattern in analyzed data and instructs the agent to treat it as a finding rather than an instruction. This effectively mitigates indirect prompt injection risks.
- [COMMAND_EXECUTION]: All identified commands (grep, git rev-parse) are used for read-only analysis of the codebase. The skill explicitly forbids mutating operations like code modification, installs, or builds. The 'execute' variant dispatches an executor subagent to an isolated worktree only after a human-reviewed plan is generated, maintaining a safe workflow.
- [DATA_EXFILTRATION]: The skill does not contain any network-facing tools or patterns that would lead to data exfiltration. All generated content is restricted to creating local documentation in the 'plans/' directory.
- [EXTERNAL_DOWNLOADS]: There are no references to external URLs, remote scripts, or unverified package installations in the provided files.
Audit Metadata