The Agent Skills Directory

[SAFE]: The skill is designed for social media strategy consultation. It utilizes a knowledge base derived from Meta patents and case studies to provide account analysis, posting schedules, and engagement strategies without requiring access to sensitive system resources.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data regarding account status, performance metrics, and content descriptions in the diagnostic and analysis modes.
Ingestion points: User-provided descriptions of account positioning, engagement statistics, and recent post history in SKILL.md and templates/post-diagnosis.md.
Boundary markers: The instructions do not define specific delimiters for separating user-provided data from system instructions.
Capability inventory: The agent's capabilities are limited to text analysis and strategy generation; it lacks access to file-writing, shell execution, or external network tools that could be abused via injection.
Sanitization: There is no explicit logic for sanitizing or escaping user input before interpolation into the analysis framework.
[COMMAND_EXECUTION]: The README.md includes a standard platform installation command claude install-skill azuma520/threads-algorithm-skill, which is the expected method for deploying the skill and poses no security risk.
[DATA_EXPOSURE]: No hardcoded credentials, API keys, or sensitive file paths were detected. The technical references to patent numbers (e.g., US10579688B2) are purely informational.

threads-algorithm