The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates multi-agent orchestration patterns, which inherently creates an attack surface for indirect prompt injection through multi-step data processing chains where one agent's output informs another's actions.
Ingestion points: Orchestrated agents (e.g., security-auditor, backend-specialist) are designed to analyze codebase data using Read, Glob, and Grep tools as described in the orchestration patterns in SKILL.md.
Boundary markers: The synthesis protocol and patterns do not explicitly provide or mandate delimiters for external data.
Capability inventory: The skill is restricted to file discovery and reading tools (Read, Glob, Grep) as defined in the allowed-tools frontmatter.
Sanitization: There are no specific instructions for sanitizing or escaping content before it is passed between specialized agents or consolidated in the synthesis protocol.
[SAFE]: No other security concerns were identified. The skill contains no executable code, remote dependencies, or obfuscated content. Its primary function is providing instructional templates for agent coordination.

parallel-agents