The Agent Skills Directory

[PROMPT_INJECTION]: The skill's debugging playbook in references/_debug-boundaries.md instructs the agent to fetch and analyze raw HTML and hydrated DOM content from the application under test (e.g., https://app.local/dashboard) using curl and mcporter. This external content is ingested into the agent's context without sanitization or boundary markers (such as "ignore embedded instructions" warnings). A malicious or compromised application could potentially influence agent behavior by embedding instructions within the HTML tags or text.\n
Ingestion points: curl output and mcporter.evaluate_script results (found in references/_debug-boundaries.md).\n
Boundary markers: Absent.\n
Capability inventory: Shell access (curl, next build, diff) and browser automation (mcporter).\n
Sanitization: Absent.\n- [EXTERNAL_DOWNLOADS]: The skill references and instructs the agent to use configuration files and related skills from an external GitHub repository (github.com/steipete/agent-scripts) which is not associated with the skill author or a trusted organization. Specifically, it points to mcporter-config.md for setup of the browser debugging tools.\n- [COMMAND_EXECUTION]: The debugging instructions in references/_debug-boundaries.md provide several shell commands for the agent to execute, including next build, curl, and mcporter tool calls. These commands are intended to inspect the build output and the rendered state of the application for debugging purposes.

nextjs-ppr-patterns