The Agent Skills Directory

[SAFE]: The skill uses environment variables (OCTAGON_API_KEY) to manage user credentials during MCP server configuration, which is a standard security best practice that avoids hardcoding secrets.
[EXTERNAL_DOWNLOADS]: The documentation provides instructions to install the Homebrew package manager using the official installation script from Homebrew's GitHub repository, which is a well-known and trusted source.
[REMOTE_CODE_EXECUTION]: The skill relies on npx to run octagon-mcp, which is the official tool provided by the vendor (OctagonAI) for this skill's functionality. This is the intended deployment mechanism for this ecosystem.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external earnings call transcripts, which could potentially contain adversarial content intended to influence the LLM's output.
Ingestion points: Earnings call transcripts referenced via ticker symbols in SKILL.md.
Boundary markers: The prompt instructions in SKILL.md do not currently specify boundary markers or delimiters for the transcript data.
Capability inventory: The skill utilizes the Octagon MCP server for data analysis.
Sanitization: No specific data sanitization or filtering logic is defined for the input transcripts.

earnings-product-pipeline