The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple sources including external URLs, file paths, and free-text ideas to drive its routing logic and tool invocations. This creates a surface where malicious instructions embedded in external content could potentially influence the agent's downstream actions.
Ingestion points: Processes user-provided URLs (GitHub, Google Docs, YouTube), local file paths, and unstructured text inputs as seen in SKILL.md.
Boundary markers: The routing logic lacks explicit delimiters or instructions for the agent to ignore commands that might be embedded within the retrieved external content.
Capability inventory: The skill uses the gh CLI to modify issues and projects, performs curl requests to an API, and invokes multiple sub-skills via the Skill tool.
Sanitization: There is no evidence of sanitization or content validation for the external data before it is passed as arguments to other tools or skills.
[COMMAND_EXECUTION]: The skill uses the gh CLI to read and update GitHub Projects and Issues, and uses ls to check for the existence of local skill definitions. These operations are performed within the vendor's (peter-tu-zynkr) own organization and repositories for administrative purposes.
[EXTERNAL_DOWNLOADS]: Fetches skill metadata and marketplace status from the vendor's official domain (https://www.zynkr.ai/api/skills) and retrieves project state from GitHub.

zynkr-skills