setup-codemap-cli
Warn
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and install a third-party tool ('codemap') from an untrusted GitHub repository ('JordanCoin/codemap') using Homebrew and Scoop. These sources are not recognized as trusted vendors.
- [PERSISTENCE]: The skill modifies agent configuration files (
settings.jsonorsettings.local.json) to register automated command execution hooks. These hooks ensure that shell commands (e.g.,codemap hook session-start,codemap hook pre-edit) are automatically executed during session initiation, before tool usage, and upon user prompt submission. - [COMMAND_EXECUTION]: The skill requires the execution of multiple shell commands to tap repositories, install software, verify the installation, and run the tool against the local filesystem.
- [INDIRECT_PROMPT_INJECTION]: The skill modifies project-level instruction files (
CLAUDE.mdorCLAUDE.local.md) to include mandatory directives (e.g., 'You MUST use codemap'). This effectively injects persistent instructions into the agent's context that will influence its behavior in all future interactions within that project.
Audit Metadata