Agent skill
execute
Use when the user wants to execute tasks autonomously, run compounder on a tasks.md file, start autonomous development, compound through a task list, or implement tasks iteratively. Guides autonomous task execution using compound loops.
Install this agent skill to your Project
npx add-skill https://github.com/jero2rome/compounder/tree/main/skills/execute
SKILL.md
Execute Skill
Orchestrates autonomous task execution using compounder's iterative loop mechanism.
Quick Start
When user has tasks ready for autonomous execution:
/compounder:compound-loop "Execute tasks from tasks.md. Mark completed with [X]. Run tests after each." --max-iterations 20 --completion-promise "ALL_TASKS_COMPLETE"
Phase Detection
Before starting, detect project state:
-
Check for tasks.md: Look for
tasks.md,.specify/specs/*/tasks.md, or similar- If found with
- [ ]uncompleted tasks: Ready for execution - Count tasks to estimate iterations (2-3 per task)
- If found with
-
Check for plan.md/spec.md: If no tasks.md exists
- Guide user to create tasks first (manually or via speckit)
Launching Compounder
Calculate Parameters
iterations = (uncompleted_tasks * 2) + 5 # buffer for debugging
completion_promise = "ALL_TASKS_COMPLETE" or custom from user
Prompt Template
Build a prompt that:
- References the tasks file explicitly
- Instructs to follow task order (phases if present)
- Mark tasks
[X]as completed - Run verification after each task
- Output
<promise>COMPLETION</promise>only when truly done
Example Invocations
For a tasks.md with 10 tasks:
/compounder:compound-loop "Execute all tasks in tasks.md following TDD. After each task: mark [X], run tests. When ALL tasks complete and tests pass, output <promise>ALL_TASKS_COMPLETE</promise>" --max-iterations 25 --completion-promise "ALL_TASKS_COMPLETE"
For a single complex task:
/compounder:run-task "Implement user authentication" --iterations 10 --done-when "AUTH_WORKING"
Speckit Integration
If speckit commands are available in the project:
| Phase | Command | Output |
|---|---|---|
| Specify | /speckit.specify |
spec.md |
| Plan | /speckit.plan |
plan.md |
| Tasks | /speckit.tasks |
tasks.md |
| Execute | /compounder:compound-loop |
Implementation |
Completion Signals
The compound loop ends when:
<promise>COMPLETION_PROMISE</promise>is output (and TRUE)- Max iterations reached
/compounder:cancel-compoundis run
Monitoring Progress
# Check current iteration
head -10 .claude/compounder-*.local.md
# View task completion
grep -E "^\- \[.\]" tasks.md
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
ideate
Use when the user has an idea to brainstorm, wants to design a feature, needs help refining a concept, says "let's ideate", or describes a vague project needing structure. Guides structured brainstorming to produce a clear feature description ready for speckit.
init
Initialize a new feature workflow with git worktree. Use when the user wants to start a new feature, create an isolated development branch, says "init worktree", or needs to set up a feature branch for compounder workflow.
handoff
Compact the current conversation into a handoff document for another agent to pick up.
edit-article
Edit and improve articles by restructuring sections, improving clarity, and tightening prose. Use when user wants to edit, revise, or improve an article draft.
scaffold-exercises
Create exercise directory structures with sections, problems, solutions, and explainers that pass linting. Use when user wants to scaffold exercises, create exercise stubs, or set up a new course section.
setup-pre-commit
Set up Husky pre-commit hooks with lint-staged (Prettier), type checking, and tests in the current repo. Use when user wants to add pre-commit hooks, set up Husky, configure lint-staged, or add commit-time formatting/typechecking/testing.
Didn't find tool you were looking for?