The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of tencentcloud-sdk-python from the official Python Package Index (PyPI). This is a well-known, official library provided by the vendor for interacting with their services.
[COMMAND_EXECUTION]: The skill executes a local script scripts/main.py which performs network operations to Tencent Cloud's official API endpoint (ocr.tencentcloudapi.com) and reads local image or PDF files for processing.
[DATA_EXPOSURE]: The script reads sensitive credentials (TENCENTCLOUD_SECRET_ID and TENCENTCLOUD_SECRET_KEY) from environment variables. This is consistent with industry best practices for secure API authentication and prevents hardcoding secrets in the source code.
[INDIRECT_PROMPT_INJECTION]: As an OCR-based tool, the skill processes external image and PDF content which could theoretically contain malicious text intended to influence the AI agent's logic when it parses the OCR output.
Ingestion points: Untrusted data enters the context through image or PDF files provided via the --image-url or --image-base64 arguments in scripts/main.py.
Boundary markers: The skill produces structured JSON output, which helps the agent distinguish between recognized text and application metadata.
Capability inventory: The script performs network requests to the vendor's API and has the capability to read local files provided by the user.
Sanitization: The skill relies on the official tencentcloud-sdk-python for data serialization and secure transmission to the API.

tencentcloud-ocr-questionmarkagent