lightcraft-raw-develop
Warn
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads a portable Windows binary from the 'storytold/lightcraft' GitHub repository and a sample RAW image from 'raw.pixls.us'.\n- [REMOTE_CODE_EXECUTION]: It facilitates the download and execution of an external 'lightcraft-cli.exe' binary package which is not from a predefined trusted organization repository.\n- [COMMAND_EXECUTION]: The skill uses 'command_run' to execute the downloaded CLI tool and also recommends using 'npx' to install associated skills from the 'full-aigc-skills' organization.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes RAW image metadata and content, creating an ingestion surface for potentially malicious instructions embedded in file data.\n
- Ingestion points: RAW files are imported via 'open_file' and metadata is queried via 'catalog.query'.\n
- Boundary markers: No explicit delimiters or safety instructions are defined to separate untrusted image content from agent commands.\n
- Capability inventory: The skill possesses the capability to execute shell commands via 'command_run' and write files to disk via 'app.export'.\n
- Sanitization: There is no evidence of metadata sanitization or validation before the data is processed by the tool or presented to the agent.
Audit Metadata