The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run various commands for code verification and scanning, including tests, type checkers, and external tools like truffler, jscpd, and FixMySlop. These actions are intended for code quality assurance and are explicitly scoped to user-authorized tasks.
[EXTERNAL_DOWNLOADS]: The skill references several external resources and tools. This includes trusted sources like Anthropic's GitHub and the TypeScript documentation, as well as third-party tools like FixMySlop (github.com/MrSpideyNihal/FixMySlop) and truffler (github.com/rayhanadev/truffler). These references are documented as optional tools for the agent to use when requested.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code (such as agent-written code or branch diffs) which serves as an ingestion point for potential indirect prompt injection. 1. Ingestion points: Code files, branch diffs, and agent-written code as described in SKILL.md. 2. Boundary markers: The instructions include a requirement to 'Verify within the task's authorization' and 'report what ran', but do not specify technical delimiters or strict sanitization for the code content being reviewed. 3. Capability inventory: The skill uses shell commands for tests, type-checking, and running scanning tools like fixmyslop and truffler. 4. Sanitization: No specific sanitization or filtering of the ingested code content is mentioned in the cleaning instructions.

anti-slop-code