The Agent Skills Directory

[SAFE]: The skill is primarily instructional, providing templates and patterns for parallelizing workflows using native agent tools. It does not perform network operations, access sensitive system files, or execute unauthorized commands.
[INDIRECT_PROMPT_INJECTION]: The skill describes a methodology for spawning sub-agents via the tool, which involves interpolating task descriptions into new prompts. This creates a surface where sub-tasks could potentially be influenced by instructions embedded in the data being processed.
Ingestion points: Input provided to the XML blocks in the Parallel Tasks template and Implementation Plan sections of SKILL.md.
Boundary markers: The skill suggests a synthesis step as a best practice for aggregating results, though no explicit delimiter-based prompt engineering is provided in the template.
Capability inventory: The skill uses the native tool for concurrent execution across directory analysis, security reviews, and general implementation tasks.
Sanitization: No specific sanitization or validation of the task prompts is performed, as the skill focus is structural.

parallel-execution