Plan swarm-ready tasks from source files into a canonical OrchPlan.
npx skills add https://github.com/tkersey/dotfiles --skill select
Please help me install this Agent Skill. Skill: select Source: https://github.com/tkersey/dotfiles/tree/main/codex/skills/select Command: npx skills add https://github.com/tkersey/dotfiles --skill select
This Skill plans swarm-ready tasks from a chosen source and emits a structured OrchPlan that can be executed across multiple agents. It ensures tasks are atomic, dependencies are explicit, and work can be scheduled in parallel where safe.
Run the select skill against a repository containing a SLICES.md or a plan-*.md to generate an OrchPlan. Review the decided waves and the decision trace to coordinate manual writeback.
High-intent search queries and answers about installing and using this skill.
To plan parallel tasks with explicit dependencies, the skill reads sources like SLICES.md or plan-N.md, normalizes task metadata, and emits a canonical OrchPlan with dependency-aware, lock-safe waves for execution across multiple agents.
An OrchPlan is a structured output that schedules atomic tasks into lock-safe, parallel execution waves. You need it when decomposing sources like bead lists into independent tasks to avoid conflicts across multiple agents.
You generate a dependency-aware execution plan by running the skill against a repository containing SLICES.md. It detects the source, applies safe auto-remediation to task metadata, and outputs scheduled waves with a decision trace.
Yes, this task planning approach uses lock-based scope to schedule tasks based on readiness, ensuring atomic tasks are scheduled in parallel only where safe and preventing execution conflicts.
Yes, you can use bd bead lists as a source. The skill reads and normalizes the bead list format, applies auto-remediation, and emits a canonical OrchPlan with optional planning pipelines to drive execution.
Auto-remediation applies safe corrections to normalize task metadata from sources like plan-N.md, but it is limited to safe fixes; you must review the emitted decision trace to coordinate manual writeback for complex issues.
tkstangImport external markdown plans into OAT's canonical plan.md format.
Ayub-KhanCoordinate a four-wave incident response with three parallel agents per wave.
fro-botCoordinate multi-agent swarms with task lists and dependency-based pipelines.
curiositechDecompose natural language tasks into DAG workflows with parallel execution.
