The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (LOW): The skill includes a shell script (references/rules/download_rules.sh) that uses curl to fetch documentation files from an external repository.
Evidence: Multiple curl commands in references/rules/download_rules.sh targeting https://raw.githubusercontent.com/vercel-labs/agent-skills/.
Context: The source repository is part of the vercel-labs organization, which is a verified trusted source. Per the [TRUST-SCOPE-RULE], this finding is downgraded to LOW.
[PROMPT_INJECTION] (LOW): As a code-analysis and optimization tool, this skill presents a surface for Indirect Prompt Injection (Category 8).
Ingestion points: The skill is designed to read and analyze user-provided React and Next.js source code (e.g., components, API routes).
Boundary markers: Absent. There are no explicit instructions or delimiters in the reference files to instruct the agent to ignore instructions embedded in the code comments or strings of the files being analyzed.
Capability inventory: Low risk. The skill is primarily informational; it does not contain executable logic (like exec or eval) that would allow an injected prompt to perform malicious actions in the agent's environment.
Sanitization: Absent. No sanitization or escaping of the ingested code content is performed.
[COMMAND_EXECUTION] (SAFE): The download_rules.sh script executes shell commands, but these are restricted to fetching documentation from trusted sources and do not involve user-controlled input.

react-best-practices