The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external source code, which creates a surface for indirect prompt injection if the code contains malicious instructions intended to influence the agent's review output.
Ingestion points: The agent reads source code files and git diffs as described in the Core Review Workflow in SKILL.md.
Boundary markers: No explicit instructions or delimiters are present to separate the analyzed code from instructions, or to warn the agent against following instructions embedded in the source code.
Capability inventory: The skill utilizes shell and batch scripts (scripts/quick-audit.sh and scripts/quick-audit.bat) to execute audit tools like npm audit, eslint, and tsc.
Sanitization: No sanitization or filtering of the ingested source code is performed before analysis.
[COMMAND_EXECUTION]: The skill provides scripts for the user to execute that perform automated security and quality audits.
Evidence: README.md and SKILL.md instruct the user to run scripts/quick-audit.sh and scripts/quick-audit.bat, which perform file searches and execute development tools such as npm, eslint, and tsc.
[EXTERNAL_DOWNLOADS]: The skill integrates with several well-known security and quality services to enhance its analysis capabilities.
Evidence: README.md and SKILL.md describe the use of established tools and services including SonarQube, CodeQL, Snyk, and Semgrep.

code-reviewer