The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to download the Habitat binary from its GitHub repository (github.com/lynx-family/habitat). The documentation correctly emphasizes security best practices by instructing users to pin specific release tags and verify the binary's integrity using SHA-256 checksums before execution.
[COMMAND_EXECUTION]: Recommends standard operations for tool installation and usage, such as chmod +x to set execution permissions on the downloaded binary and ./hab sync to perform dependency synchronization. These commands are necessary for the tool's intended purpose.
[DYNAMIC_EXECUTION]: Documents the tool's action dependency type, which allows for the execution of Python functions during synchronization. The skill provides a benign example of this functionality for filesystem operations like directory creation and file copying. Users should review any custom actions in their manifests.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided configuration files (DEPS and .habitat) that could contain malicious instructions if obtained from untrusted sources.
Ingestion points: Project manifests (DEPS, .habitat) described in SKILL.md.
Boundary markers: None; the DEPS format allows Python code to be mixed with data.
Capability inventory: The tool executes Python functions and performs file operations as described in the documentation.
Sanitization: No automated sanitization is described; the tool executes the logic defined by the project owner in the manifest.

habitat-usage