The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill implements security best practices, including setting file permissions to 0o600 for sensitive configuration files.
[COMMAND_EXECUTION]: The skill executes bundled Python scripts (scripts/anygen.py, scripts/auth.py, scripts/fileutil.py) to manage authentication, poll task status, and download results. These operations are limited to the skill's own directory and authorized workspace.
[EXTERNAL_DOWNLOADS]: The skill interacts with the vendor's API at www.anygen.io and the well-known service open.feishu.cn for platform integration. These connections are necessary for the skill's primary functionality.
[DATA_EXFILTRATION]: User-provided files are uploaded to the AnyGen API for processing. The instructions explicitly require the agent to obtain user consent before reading or uploading any local data.
[PROMPT_INJECTION]: The skill processes untrusted user data from reference files to analyze slide requirements. This creates an indirect prompt injection surface.
Ingestion points: Phase 1 (reading user-specified reference files in SKILL.md).
Boundary markers: None identified in the prompt templates.
Capability inventory: sessions_spawn for background monitoring, network operations to anygen.io, and file writing to ~/.openclaw/workspace/.
Sanitization: The instructions mitigate risk by directing the agent to summarize key points rather than directly interpolating raw file content.

anygen-slide