The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides installation commands for the Blaxel CLI that fetch and execute scripts directly from the vendor's official domain.
Evidence: curl -fsSL https://blaxel.ai/install.sh | sh and irm https://blaxel.ai/install.ps1 | iex in SKILL.md.
[COMMAND_EXECUTION]: The skill facilitates the execution of arbitrary commands within remote sandbox environments as a core feature of its resource management functionality.
Evidence: Documentation in references/run.md details how to execute shell commands and scripts using bl run sandbox with various payload configurations.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources such as sandbox logs, application outputs, and agent chat responses, which constitutes an attack surface for indirect prompt injection.
Ingestion points: Data is ingested through commands like bl logs, bl chat, and bl run (found in references/logs.md, references/chat.md, and references/run.md).
Boundary markers: No explicit boundary markers or "ignore embedded instructions" warnings are present in the provided documentation files.
Capability inventory: The skill has the capability to execute shell commands, manage files, and perform network operations via the bl CLI.
Sanitization: No sanitization or escaping of external content is mentioned in the static documentation.
[EXTERNAL_DOWNLOADS]: The documentation describes functionality for downloading and refreshing agent skills from the vendor's public GitHub repository.
Evidence: The bl skills install command is documented in references/skills.md as fetching content from github.com/blaxel-ai/agent-skills.

blaxel-cli