The Agent Skills Directory

[SAFE]: The skill operates entirely on text processing and local configuration. It reads platform-specific rules from references/platform-rules.md and project-specific voice guidelines from social-context.md. No external network requests, credential harvesting, or unauthorized file system operations were detected.
[PROMPT_INJECTION]: The skill ingests untrusted user data (blog posts, notes, topics) in Step 2 of its workflow. However, it lacks exploitable capabilities such as code execution (eval/exec), shell access, or network connectivity. The risk of indirect prompt injection is mitigated by the skill's specific focus on drafting and formatting social media content without executing any logic contained within the input data.
[COMMAND_EXECUTION]: There are no shell commands, subprocess calls, or dynamic execution patterns in the provided files. The skill relies on natural language instructions for the agent to perform drafting and character counting tasks.

social-thread