The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references several external resources and assets from GitHub repositories. This includes a runtime binary for the LightCraft CLI tool (v0.8.0) hosted at github.com/storytold/lightcraft and documentation/workflow facts from github.com/darktable-org/dtdocs and github.com/dpnkrpl/awesome-lightroom-presets. These are used for tool configuration and as reference material for the AI agent.
[COMMAND_EXECUTION]: The skill is designed to interact with the host system using the command_run tool to invoke the lightcraft-cli.exe. It executes specific commands such as catalog.query, library.select, develop.set, and app.export to perform photo editing tasks. The instructions emphasize verifying actual tool schemas and command availability before execution.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted external data.
Ingestion points: The skill reads file metadata through catalog.query and imports external image files via open_file and CLI import commands.
Boundary markers: None. There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content embedded in image metadata or EXIF tags.
Capability inventory: The skill possesses the ability to execute shell-level commands via command_run and write files to the local file system using the app.export command.
Sanitization: Not detected. The provided instructions do not specify any validation or sanitization steps for data retrieved from external image files or their metadata.

lightcraft-portrait-develop