The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs the agent to execute a setup script directly from a remote URL using a shell pipe pattern.
Evidence: bash <(curl -fsSL https://payram.com/setup_payram_agents.sh) --testnet in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill fetches deployment plans and operational recipes from a remote Model Context Protocol (MCP) server.
Evidence: Interaction with https://mcp.payram.com/mcp to retrieve ordered setup steps via payram_setup_plan.
[COMMAND_EXECUTION]: The instructions involve running various CLI tools and scripts to manage blockchain operations, including wallet creation and worker management.
Evidence: Usage of payram_doctor, payram_runbook, and payram_ops_playbook tools.
[INDIRECT_PROMPT_INJECTION]: The skill relies on output from external tools (MCP, blockchain nodes, and local workers) to determine its next actions, creating an ingestion surface for untrusted data.
Ingestion points: payram_setup_plan output from the hosted MCP, and balance/sync data from local or remote blockchain nodes.
Boundary markers: None explicitly defined in the markdown instructions to differentiate tool output from agent instructions.
Capability inventory: The skill allows file system interaction, shell command execution (installation), and network requests (node queries).
Sanitization: No evidence of sanitization or validation of data returned by the payram_ops_playbook or MCP tools before the agent processes the results.

payram-agent-journey