The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The SKILL.md file provides installation instructions that use a piped shell execution pattern: curl -fsSL https://install.iii.dev/iii/main/install.sh | sh. This pattern downloads and executes code from an external server without a prior review step.
[COMMAND_EXECUTION]: The skill documents the use of the modules::shell::ExecModule within the engine's configuration (iii-config.yaml). This module is designed to run arbitrary shell commands sequentially during the engine's startup process.
[EXTERNAL_DOWNLOADS]: The skill requires downloading external assets including the iiidev/iii Docker image from Docker Hub and the iii-sdk package from npm.
[INDIRECT_PROMPT_INJECTION]: The skill defines a system that ingests external data from multiple sources (HTTP requests, queue messages, and log entries) and processes them through user-defined function handlers.
Ingestion points: Untrusted data enters the agent context through ApiRequest objects (in SKILL.md), queue messages, and log entries.
Boundary markers: There are no specified delimiters or warnings to ignore embedded instructions in the processed data.
Capability inventory: The skill possesses capabilities for file system writes via State/KV modules (in references/engine-modules.md), network operations via the call function (in SKILL.md), and arbitrary subprocess execution via ExecModule (in references/engine-modules.md).
Sanitization: The documentation does not specify sanitization or validation logic for data entering the function handlers.

iii