cursor-env-setup
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEDATA_EXFILTRATIONPRIVILEGE_ESCALATIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The script
scripts/git-identity.shaccesses and writes to the sensitive file path~/.ssh/git_signing_key. This is used to store the user-provided private SSH key for commit signing. While this is the intended purpose, it involves handling raw private key material. - [PRIVILEGE_ESCALATION]: The skill uses
chmod -xto disable executable scripts within the directory specified by thecore.hooksPathgit configuration. This is done to neutralize default platform hooks (e.g.,*.co-authorhooks) so they do not interfere with the custom identity setup. Modifying execution permissions on files in platform-managed directories is a form of privilege adjustment. - [INDIRECT_PROMPT_INJECTION]: The git hook scripts process external, potentially untrusted data such as git commit messages and branch names.
- Ingestion points: The
commit-msghook reads the commit message file, and thepre-commithook reads the current branch name. - Boundary markers: None identified.
- Capability inventory: The skill has the ability to modify git configuration, change file permissions, and rewrite commit messages.
- Sanitization: The skill uses
awkfor basic string manipulation and pattern matching on commit messages. - [COMMAND_EXECUTION]: The skill documentation and templates instruct the agent to make scripts executable using
chmod +xand to run setup scripts using thebashinterpreter, which is standard for automation tasks.
Audit Metadata