kata-jido-agent
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and analyze potentially untrusted content from the local repository, including
mix.exs, source code, and project-specific instructions. This creates an attack surface where malicious data within these files could attempt to redirect the agent's behavior. - Ingestion points:
SKILL.md(instructions to readmix.exs, nearby agents, actions, plugins, and supervision code). - Boundary markers: None specified in the instructions.
- Capability inventory: The skill involves generating Elixir code and executing shell-based "documented checks" and "focused tests" (e.g., via
mix test). - Sanitization: No explicit sanitization or validation of the ingested repository content is mentioned.
- [COMMAND_EXECUTION]: The skill relies on executing shell commands to perform verification tasks and run repository tests. While expected for a development tool, this capability increases the impact of other vulnerabilities such as indirect prompt injection.
Audit Metadata