The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes output from static analysis tools and reads local rule definitions to guide code modifications, which creates a potential surface for indirect instructions to influence the agent. (1) Ingestion points: The agent reads JSON output from oxlint commands and rule implementation files located in tools/oxlint/anti-slop/rules/. (2) Boundary markers: No explicit delimiters or boundary instructions are present to distinguish tool output from agent instructions. (3) Capability inventory: The skill uses shell command execution via bunx and bun run as well as file system access for code modification. (4) Sanitization: No specific validation or sanitization of tool output is performed before it is processed by the agent.
[COMMAND_EXECUTION]: The skill uses shell commands to interact with the repository, including linting, type-checking, and running tests. Evidence: Instructions include running bunx oxlint, bunx tsc, and various bun run commands to verify the migration progress.

anti-slop-migration