The Agent Skills Directory

[COMMAND_EXECUTION]: The skill directs the agent to execute shell commands for running tests, specifically mentioning pnpm and vitest. It also utilizes repository-specific scripts such as scripts/run-vitest-stable.mjs and scripts/general-server-shard.mjs to analyze test performance and validity.
[INDIRECT_PROMPT_INJECTION]: The skill creates an exposure to indirect prompt injection by analyzing untrusted files from the repository while having the ability to execute commands and write to the filesystem.
Ingestion points: The agent is instructed to read test source files, package manifests (package.json), and CI configurations (.github/workflows/pr.yml).
Boundary markers: There are no instructions provided to use specific delimiters or protective prompts (e.g., 'ignore previous instructions') when reading the content of these files.
Capability inventory: The agent has permissions to execute shell commands, edit source code files, and manage pull requests.
Sanitization: The instructions do not define any validation or sanitization steps for the content of the files being audited before they are processed by the agent.

test-audit