The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is designed to process untrusted code from Pull Requests which could contain malicious instructions targeting the agent.
Ingestion points: Processes repository files via read_file during PR audits.
Boundary markers: Lacks explicit boundary markers or instructions to ignore embedded commands within the code being reviewed.
Capability inventory: Possesses run_command, write_file, and edit_file tools which could be abused if an injection is successful.
Sanitization: No evidence of input sanitization or validation for processed code files.
[Command Execution] (LOW): The skill uses the run_command tool to perform automated checks such as linting and testing.
Evidence: SKILL.md identifies run_command as a tool and the 'Operational Sequence' includes executing CI checks (Lint, Test, Build).
Context: While this is a primary function of the skill, executing scripts (e.g., npm run lint) on untrusted branches carries an inherent risk of arbitrary code execution via compromised build configurations.

resonance-reviewer