The Agent Skills Directory

[METADATA_POISONING]: The skill documentation in SKILL.md and references/build-plan-template.md includes references to hypothetical and currently non-existent AI model versions, such as 'Gemini 2.5 Pro', 'Claude Sonnet 4.5', and 'GPT-5.2 Codex'. This is factually inaccurate and deceptive concerning the required infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface for indirect prompt injection by instructing the agent to process data from unverified sources. • Ingestion points: The agent audits potentially untrusted user asset descriptions (references/archive-audit.md) and extracts data from community forums like Reddit and Discord during the market signal scan (references/market-signal-scan.md). • Boundary markers: No delimiters or protective instructions are provided to help the agent distinguish between data inputs and operational instructions. • Capability inventory: The skill has permission to write to Google Drive and use the 'manus-upload-file' command to deploy web content. • Sanitization: There are no mechanisms described for sanitizing or validating external content before it is used to generate deliverables or code templates.
[COMMAND_EXECUTION]: The skill instructs the agent to use a non-standard CLI tool 'manus-upload-file --webdev' for image hosting and deployment. While this appears to be a platform-specific utility for the skill's primary purpose, it involves executing shell commands with parameters derived from the design process.

digital-product-inventor