The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill utilizes Firecrawl to extract brand identity data (typography, tone, and raw markdown) from arbitrary user-provided URLs to influence slide generation.
Ingestion points: The extract_brand_dna function and the general brand extraction logic ingest untrusted data from external websites into the agent's context.
Boundary markers: The instructions lack explicit delimitation or 'ignore embedded instructions' warnings for the data retrieved from external URLs.
Capability inventory: The skill has the capability to write files (slides.html) and perform network operations via the Firecrawl API.
Sanitization: There is no documented evidence of sanitization or validation of the content scraped from external websites before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill's installation instructions direct the user to clone a repository from a non-trusted GitHub account (github.com/ItsssssJack/power-design).
[COMMAND_EXECUTION]: Documentation for 'Advanced Usage' and 'API Integration' includes Python and Bash code examples that utilize subprocess.run to invoke the claude CLI. While presented as user-facing examples, this pattern demonstrates a capability for command-line execution that requires careful handling to prevent command injection if adapted for automated tasks.

power-design-slides