The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data by reading and grepping site source code, footer information, and contact pages. This creates a surface for indirect prompt injection where malicious instructions embedded in the analyzed web content could attempt to influence the agent's behavior.
Ingestion points: The agent extracts information from site files using Grep and Read tools as described in §6 and §7 of SKILL.md.
Boundary markers: The skill does not provide specific instructions to use delimiters or ignore commands found within the processed files.
Capability inventory: The skill has access to Bash, Write, and Edit tools to perform its tasks.
Sanitization: No explicit sanitization or filtering of the ingested site content is mentioned.
[COMMAND_EXECUTION]: The skill requires the execution of shell commands through pnpm and python to run automated auditing tools (§13.1, §13.5.3). These tools interact with the local environment to generate SEO reports and performance metrics.

web-local-seo