The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines a 'Memory Protocol' that requires the agent to execute the bash command cat .claude/context/memory/learnings.md to retrieve context.
[PROMPT_INJECTION]: The instructions use high-priority directives such as 'Iron Laws', 'MANDATORY', and 'NEVER' to constrain model behavior, which are common patterns in prompt injection.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to analyze untrusted external data (React code).
Ingestion points: The skill reviews files matching **/*.tsx and **/*.jsx via the Read and Glob tools.
Boundary markers: There are no explicit delimiters or warnings defined to prevent the agent from following instructions embedded in the analyzed code.
Capability inventory: The skill environment provides the agent with powerful tools including Bash, Write, Edit, and Grep.
Sanitization: No sanitization or validation of the input code is performed before analysis.

react-expert