The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to execute several Python scripts to perform validation, site management, and rendering tasks. The scripts, including validate_query_corpus.py, validate_seo_findings.py, manage_markdown_site.py, manage_markdown_work.py, and render_keyword_map.py, are expected to be present in a scripts/ directory within the suite's root. The path to these scripts is dynamically determined via the variable, which is resolved from .seo-suite-runtime.json or environment variables.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, untrusted sources, creating a potential surface for indirect prompt injection. 1. Ingestion points: Untrusted data enters the agent's context from search engine result pages (SERPs), competitor websites, and general web discovery during the research workflow (as specified in SKILL.md and keyword-research-protocol.md). 2. Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' markers when processing external content. 3. Capability inventory: The skill provides the agent with capabilities to execute shell commands (running Python scripts) and write to the file system (SKILL.md, references/markdown-delivery-contract.md), which could be targeted by an injection attack. 4. Sanitization: The protocols do not describe any mechanisms for sanitizing or escaping the text retrieved from external sources before it is processed by the agent.

seo-research