repo-rules-reader
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill implements a workflow that ingests content from multiple repository files (such as
AGENTS.mdanddocs/*) and instructs the agent to apply these rules to all subsequent tasks. This behavior creates a surface for indirect prompt injection where an attacker who can modify repository files could redirect the agent's behavior. The skill lacks boundary markers or explicit instructions to ignore potentially malicious content within these files. - [COMMAND_EXECUTION]: Instructions in
SKILL.mddirect the agent to identify and include a shell command string (source source_me.sh && python3) derived from the contents ofAGENTS.md. This facilitates the execution of local scripts, creating a dependency on the integrity of the repository's file system.
Audit Metadata