The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external artifacts, code, and user briefs for readiness reviews. This creates a surface where malicious instructions could be embedded in the content being reviewed. While inherent to the skill's purpose as an analysis tool, the instructions mitigate this by directing the agent to identify failures and distinguish real behavior from mocks.
Ingestion points: SKILL.md instructs the agent to identify artifacts, code, and user context for review.
Boundary markers: No specific delimiters or safety warnings are required for the untrusted content being processed.
Capability inventory: The instructions suggest the agent should "probe the endpoint directly", implying network interaction capability if tools are available.
Sanitization: No explicit sanitization or filtering of reviewed content is defined before assessment.
[SAFE]: The skill instructions provide a methodology for security and readiness reviews without introducing dangerous behaviors. It correctly identifies common pitfalls like relying on hidden UI for security and explicitly forbids the agent from publishing, messaging stakeholders, or altering permissions as a side effect of the review process.

readiness