The Agent Skills Directory

[NO_CODE]: The skill is composed entirely of markdown instructions and YAML metadata. It does not include any executable scripts, binaries, or configuration files.
[COMMAND_EXECUTION]: The instructions mention the capability to use standard security and static analysis tools such as npm audit, pip-audit, Snyk, and Semgrep. These references describe the agent's domain expertise and do not represent hidden or malicious command execution.
[EXTERNAL_DOWNLOADS]: The skill references well-known technology platforms and security services including GitHub, GitLab, SonarQube, and Snyk. These are standard references for the intended use case and do not involve downloading untrusted remote code.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted data in the form of code changes for review, which presents a surface for indirect prompt injection.
Ingestion points: The skill processes code snippets, pull request content, and configuration files provided during the review process.
Boundary markers: No specific delimiters or "ignore instructions" markers are defined in the response instructions for handling code content.
Capability inventory: The skill describes capabilities for running security audits and static analysis tools if the environment permits.
Sanitization: There are no explicit instructions for the AI to sanitize or escape potentially malicious instructions embedded within the code it reviews.

code-reviewer