The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires cloning a repository from a personal account (https://github.com/pennydinh/marketing-pineline-share.git) and running npm install. The source is not a recognized trusted vendor or organization.- [INDIRECT_PROMPT_INJECTION]:
Ingestion points: The researchTopic function in SKILL.md (Core Workflow 1) fetches content from external sources including Twitter/X, LinkedIn, and TechCrunch.
Boundary markers: Absent. The fetched research data is directly stringified into the LLM prompt using ${JSON.stringify(research)} without delimiters.
Capability inventory: The skill performs file system writes via renderMedia (./output/${content.title}.mp4) and manages API keys for external services.
Sanitization: No sanitization, filtering, or escaping logic is applied to the external research data before it is passed to the AI models.- [COMMAND_EXECUTION]: The installation and troubleshooting sections instruct the agent to execute shell commands such as npm install, npm run build, and renderMedia which invokes Chromium with the --no-sandbox flag.

marketing-content-pipeline-automation