The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various CLI tools including git, uv, npm, node, greptile, coderabbit, and the GitHub CLI (gh).
It runs npm run lint, npm run build, and npm run validate, which executes arbitrary scripts defined in the local project's package.json file.
It uses uv run lintro to perform linting checks.
It executes greptile review and coderabbit review for automated code auditing.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted extension code and metadata (e.g., files in src/, package.json, CHANGELOG.md).
Ingestion points: Reads local project files and code files using cat, rg, and node.
Boundary markers: No explicit instructions are provided to the agent to ignore or isolate embedded instructions within the project files.
Capability inventory: Includes the ability to execute shell commands, write to the filesystem (via git), and interact with the network (via gh and external review tools).
Sanitization: There is no evidence of sanitization for data extracted from project files before it is used in the PR body or other command arguments.
[EXTERNAL_DOWNLOADS]: The skill utilizes npm and uv, which are package managers that typically download dependencies from remote registries.
It integrates with external services Greptile and CodeRabbit for code review, which involves transmitting local code to these platforms.

pr-raycast