The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted data from multiple sources including materialText, materialUrl, and materialFile (defined in SKILL.md and references/form-schema.json).
Ingestion points: Data enters the agent via user-provided text, file uploads, and external URL references.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are provided within the skill's instruction set.
Capability inventory: The skill's runner script (scripts/run.py) performs network requests to the vendor API but does not execute local system commands based on AI output.
Sanitization: No client-side sanitization or validation of external content is performed before it is sent to the API.
[EXTERNAL_DOWNLOADS]: The skill's Python runner (scripts/run.py) performs network operations targeting the vendor's official API at https://ai-skills.ai to process the diagnosis logic.
[COMMAND_EXECUTION]: A Python script scripts/run.py is provided to allow users to interact with the API. This script utilizes standard Python libraries and performs authorized network communication with the vendor's infrastructure.

agent-browser