The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructs the agent to store the user's API key in ~/.pixazo/api-key. While the instructions include setting restricted permissions (chmod 600), storing sensitive credentials in a predictable local file creates a target for local data exposure.
[EXTERNAL_DOWNLOADS]: The skill is instructed to fetch a remote markdown file from https://www.pixazo.ai/models/pixelforge.md to determine exact parameter names and accepted values during execution. This represents a dependency on external content to define the agent's logic.
[DYNAMIC_EXECUTION]: The documentation mentions the use of npx skills add Pixazo-AI/skills --skill '*', which involves the installation and potential execution of remote packages from a third-party repository.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by processing external data and user prompts.
Ingestion points: Data enters the context via user-supplied prompts, reference image data, and the remote content fetched from https://www.pixazo.ai/models/pixelforge.md.
Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions embedded within the ingested data.
Capability inventory: The skill uses network operations (curl, requests, fetch) and file system operations (writing to ~/.pixazo/api-key).
Sanitization: There is no mention of escaping, validation, or filtering of the external content before it is processed by the agent.

pixelforge