The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection by ingesting untrusted data from multiple sources and using it to generate content and command parameters.
Ingestion points: The skill reads repository content using tools like grep and jq and performs up to three web searches and page fetches to gather facts (Step 1).
Boundary markers: There are no instructions to use delimiters or protective markers to prevent the agent from obeying instructions embedded in the gathered data.
Capability inventory: The skill uses python to execute assembly and preview scripts, runs shell commands for data extraction, and writes several files to the local file system (Steps 3 and 4).
Sanitization: No sanitization or validation steps are defined for the data retrieved from the repository or external web pages before it is processed.
[COMMAND_EXECUTION]: The skill automates the execution of several shell commands and Python scripts.
Evidence: It explicitly instructs the use of grep -c and jq for fact-finding and executes a chain of Python scripts (preview.py, assemble.py, check.py) located within an external explainer-kit path to build the final output.

explain-visually