The Agent Skills Directory

[PROMPT_INJECTION]: Indirect prompt injection surface exists through the ingestion of untrusted data from the Edge Impulse platform.
Ingestion points: The skill instructions involve fetching job logs from /jobs/{jobId}/stdout/download and sample metadata or lists from the Studio API.
Boundary markers: Absent; the instructions do not define delimiters to separate agent system prompts from the content of fetched external data.
Capability inventory: The skill facilitates shell command execution and network operations through the Edge Impulse CLI suite (e.g., edge-impulse-daemon, edge-impulse-uploader, edge-impulse-linux-runner).
Sanitization: Absent; the skill directs the agent to display error lines from downloaded job logs without explicit sanitization or filtering logic.
[COMMAND_EXECUTION]: The skill documentation facilitates the use of official Edge Impulse command-line utilities for platform automation and hardware interaction.
Evidence: Documentation and recipes for edge-impulse-daemon, edge-impulse-uploader, edge-impulse-run-impulse, and edge-impulse-linux-runner.
[EXTERNAL_DOWNLOADS]: The skill references the installation of platform tools and libraries from established package registries and official repositories.
Evidence: Instructions for installing edge-impulse-cli and edge-impulse-linux via npm, and references to official GitHub repositories within the edgeimpulse organization.

api