The Agent Skills Directory

[SAFE]: The skill implements proactive security measures, such as explicit instructions to redact secrets before passing context to sub-agents as documented in references/context-pack.md.- [EXTERNAL_DOWNLOADS]: The skill references an external output schema from the author's GitHub repository (https://raw.githubusercontent.com/csark0812/toolbox/main/.skeleton/references/output-schema.md). This is a documented vendor resource used for formatting and validating consolidated report outputs.- [PROMPT_INJECTION]: The skill orchestrates data across multiple agents, creating an inherent attack surface for indirect prompt injection.
Ingestion points: The coordinator processes external artifacts, user prompts, and responses from various sub-agents (e.g., as described in SKILL.md and references/output-format.md).
Boundary markers: Structured headers such as ### Context pack and 'Constraints' blocks are used to delimit sub-agent instructions and separate different slices of work.
Capability inventory: The skill manages the invocation of Task and Subagent tools to perform delegated logic.
Sanitization: The skill includes explicit directives to redact sensitive information before it is passed into member prompts and enforces context asymmetry to prevent agents from seeing unauthorized execution history.

subagents