The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the agent-browser tool from the official GitHub repository of a well-known and trusted organization (vercel-labs). This is used to enhance the agent's research capabilities and is documented with standard installation procedures.
[COMMAND_EXECUTION]: The skill utilizes shell commands to verify tool availability and perform browser-based research tasks. These commands include npx, npm, and agent-browser, which are necessary for its primary function of competitive analysis.
[INDIRECT_PROMPT_INJECTION]: The skill has a defined attack surface for indirect prompt injection as it is designed to ingest and process data from external websites (competitor pricing pages, review platforms like G2, and social media) via the agent-browser tool.
Ingestion points: External content is retrieved into the agent's context through commands like agent-browser get text body as seen in references/competitive-research.md.
Boundary markers: The workflow relies on manual logical verification phases rather than explicit technical delimiters between untrusted web data and internal instructions.
Capability inventory: The skill possesses capabilities for file system writes to generate collateral and shell execution for research tools.
Sanitization: The skill incorporates a specific 'Phase 4: Validation' workflow to verify the accuracy and source of all claims before they are included in final deliverables.

paw-mkt-sales