The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructs the agent to manage and utilize sensitive Twitter OAuth1 credentials (consumer key, secret, access token, and access secret) and Buffer API tokens stored in local status files like MASTER-STATUS.md.
[INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting content from external sources such as podcasts, interviews, and blog posts to populate a 'Source Library' for tweet generation, which serves as an attack surface for indirect prompt injection.
Ingestion points: Data is pulled from external transcripts and articles into SOURCE-INDEX.md and other markdown files.
Boundary markers: The agent uses markdown table delimiters and templated sections to separate data.
Capability inventory: The agent can post content to external APIs (Twitter, Buffer) and write to the local filesystem.
Sanitization: The instructions suggest manual verification of facts, but there is no programmatic escaping or sanitization of external content before it is interpolated into generating prompts.
[DATA_EXFILTRATION]: The skill is designed to send information to external services (Twitter/X and Buffer APIs) as part of its core functionality.
[EXTERNAL_DOWNLOADS]: The documentation provides installation instructions for components hosted on the author's GitHub repository (Gingiris-1031), HuggingFace (Gingiris), and official vendor domains.

gingiris-twitter-agent-ops