photocraft-cli-project

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads the PhotoCraft CLI and an optional macOS desktop application from the project's official repository on GitHub. These downloads are required for the skill's primary functionality and are sourced from a well-known platform.
  • [COMMAND_EXECUTION]: It utilizes Python scripts (such as bootstrap.py, cli.py, and workflow.py) provided by the photocraft-harness dependency to manage the runtime environment and execute native image processing commands.
  • [DYNAMIC_EXECUTION]: The skill constructs and executes image manipulation workflows based on structured JSON plans (craft-command-plan/v1). This mechanism allows the agent to perform complex editing tasks through a series of defined operations.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill reads and processes external .pcraft project files and various image assets (PNG, PSD, etc.) provided by the user.
  • Boundary markers: It employs strict JSON schema validation for all command plans and operations, which helps define clear boundaries for the agent's actions.
  • Capability inventory: The skill possesses the capability to read/write files and execute the PhotoCraft CLI via subprocesses to perform image edits.
  • Sanitization: All inputs are parsed through structured JSON handlers that enforce strict key uniqueness and type checks, minimizing the risk of processing malicious instructions embedded in project metadata.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 01:36 PM