The Agent Skills Directory

[METADATA_POISONING]: The skill's YAML frontmatter identifies the author as 'OpenAEC-Foundation', which is inconsistent with the expected author context 'impertio-studio'. This discrepancy in metadata can mislead users or automated systems regarding the skill's true provenance.
[INDIRECT_PROMPT_INJECTION]: The skill provides numerous templates and patterns for ingesting untrusted data from external sources that are then processed by the agent's context.
Ingestion points: Code snippets in SKILL.md (Pattern 2, 5, 6) and references/examples.md (Example 3, 4, 11, 12) facilitate reading external file contents, network data, and IPC bodies.
Boundary markers: The instructions and code examples do not include boundary markers or delimiters to isolate untrusted data from the agent's primary instruction set.
Capability inventory: The skill explicitly documents and enables capabilities for filesystem access (reading and writing via tokio::fs and std::fs) and network operations (via reqwest).
Sanitization: The provided examples lack proper input validation. Specifically, 'Example 11' in references/examples.md demonstrates writing binary data to a file path constructed directly from an unsanitized request header (x-filename), which represents a classic path traversal vulnerability.

tauri-syntax-commands