The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill directs the agent to ingest untrusted data by analyzing the "top 5-10 ranking pieces" from the web during competitive research (Mode 1, Step 1). This creates an attack surface where malicious content on external websites could attempt to influence the agent's behavior or output. * Ingestion points: External web pages accessed for competitive analysis. * Boundary markers: Absent; the instructions do not specify delimiters or safety warnings for processing external content. * Capability inventory: The agent writes draft files and executes local Python scripts (scripts/content_scorer.py) for analysis. * Sanitization: Absent; there is no instruction to validate or filter the content retrieved from external sources.- [COMMAND_EXECUTION]: The skill instructions require the agent to execute local scripts (e.g., scripts/content_scorer.py) to perform readability and quality checks on generated content. While these scripts are part of the skill and perform benign operations, the workflow relies on the execution of local code.

content-production