framer-agent-playbook
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill is configured to fetch updates and reference content from the author's public GitHub repository (
fredm00n/framerlabs). This allows the agent to maintain an up-to-date knowledge base of Framer API changes and development traps. - [REMOTE_CODE_EXECUTION]: Implements a mechanism where the agent adopts remote instructions (
SKILL.md) if a newer version is identified upstream. In the context of an AI agent, replacing instructions constitutes dynamic behavior modification, which is here restricted to the identified vendor's repository. - [COMMAND_EXECUTION]: Instructs the agent to use various system commands such as
git clonefor version management andnpxutilities for setting up the Framer environment, running code component tests, and managing project dependencies. - [PERSISTENCE]: If the user adds
auto-updateto their project configuration, the skill enables the agent to overwrite its own local files with the latest content from the remote repository, ensuring persistent alignment with the upstream source. - [INDIRECT_PROMPT_INJECTION]: Since the skill processes external project data including CMS collections and node structures, it is exposed to potential injection from project metadata. The skill mitigates this by providing explicit instructions for the agent to sanitize all client names, keys, and tokens before any community-driven data sharing.
Audit Metadata