The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The AutoGPT platform described in this skill is designed to ingest data from various untrusted sources, such as webhooks and external integrations including GitHub, Google, Discord, and Notion. This data is subsequently processed by Large Language Models (LLMs) which may then trigger autonomous actions based on the content.
Ingestion points: Webhook triggers and integration blocks for third-party services are defined as entry points in SKILL.md.
Boundary markers: The documentation does not specify the use of delimiters or explicit instructions for the AI to ignore embedded commands within the processed data streams.
Capability inventory: The platform possesses significant capabilities including making network requests, performing database operations, and executing custom code blocks as detailed in references/advanced-usage.md.
Sanitization: No specific sanitization, escaping, or validation logic for external content is described in the provided skill files.
[DYNAMIC_EXECUTION]: The platform supports the development and runtime execution of custom Python blocks and agents. This is a core functionality of the Forge toolkit and the visual builder described in references/advanced-usage.md, allowing for the execution of user-defined logic within the platform's backend environment.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to clone the AutoGPT repository from its public GitHub repository and install dependencies using standard package managers such as npm and poetry.

autogpt-agents