The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from multiple sources. \n
Ingestion points: User input via the $ARGUMENTS placeholder, architecture documentation (data-model.md), and cross-project memory (failure_patterns.json). \n
Boundary markers: The skill lacks explicit delimiters or instructions to ignore embedded commands within the ingested data. \n
Capability inventory: The skill has the ability to execute shell commands (npm test), modify project files, and invoke other agent skills (executeCommand). \n
Sanitization: There is no evidence of robust sanitization or validation of external inputs before they are used to influence the agent's hypothesis generation or fix implementation. \n- [COMMAND_EXECUTION]: The skill utilizes dynamic context injection to execute the system date command (!date) at load time. This is a benign use case intended to provide a timestamp for the diagnostic process. \n- [COMMAND_EXECUTION]: The skill programmatically invokes development-related shell commands such as npm test and npm run build, as well as internal skill commands like executeCommand('data-model'), as part of its core diagnostic and repair workflow.

fix