skills.rest
wesbragagt
wesbragagtCOMMUNITY
Explore all skills by this author →
Website𝕏@wesbragagtGitHub

code

Executes tasks from a tasks.yaml file by orchestrating code-writer and verifier agents.

1|Updated May 4, 2026
Search Tags:#code generation#verification#spec-driven development#task orchestration#dependency resolution#tasks.yaml#subagent delegation
One-click install
npx skills add https://github.com/wesbragagt/nixos-config --skill code-wesbragagt
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: code
Source: https://github.com/wesbragagt/nixos-config/tree/main/home/skills/code
Command: npx skills add https://github.com/wesbragagt/nixos-config --skill code-wesbragagt

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Coordinating multi-step implementation work from a task plan is error-prone: dependencies get missed, tasks run in the wrong order, and completed work goes unverified. This Skill automates execution of a tasks.yaml file by resolving dependencies, dispatching implementation agents, and gating completion on independent verification. ## Core Features & Use Cases - Dependency-aware orchestration: Reads tasks.yaml, finds ready tasks whose dependencies are done, and runs them sequentially or in parallel (up to 3 concurrent agents) based on file overlap. - Agent delegation with verification: Dispatches code-writer-simple or code-writer-complex agents with a structured task packet, then requires an independent verifier subagent to return an evidence-backed PASS before marking any task done. - Status tracking via CLI: Uses tasks.py commands (summary, ready, set, verify) to keep task state consistent and report progress. - Use Case: You have a spec broken down into .specs/auth-feature/tasks.yaml. Run the Skill against that file and it executes each task in dependency order, verifying each implementation before moving on. ## Quick Start Ask the agent to run the tasks in .specs/auth-feature/tasks.yaml and execute each ready task with verification until all are done.

Frequently Asked Questions about code

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I execute a tasks.yaml file with AI agents?▼

Run the code skill with the path to your tasks.yaml file. It loads the task summary, finds ready tasks whose dependencies are done, dispatches implementation agents, and marks tasks done only after an independent verifier returns PASS.

How do I run only one specific task from tasks.yaml?▼

Pass the --task flag with the task key, for example /code ./tasks.yaml --task setup-database. This jumps directly to executing that task and skips dependency resolution for other tasks.

When does the skill run tasks in parallel versus sequentially?▼

Tasks run in parallel (maximum 3 concurrent agents) only when they touch completely separate files or domains. Tasks that may modify the same files, share configuration, or have overlapping paths run sequentially.

What happens if the verifier returns FAIL or PARTIAL?▼

The task stays in progress status and is never marked done. The verifier's evidence and verdict are reported back so you can review the issues and retry the task.

Why are there no ready tasks but the plan is not complete?▼

This means remaining tasks are blocked by unfinished dependencies, stuck in progress, or part of a circular dependency. The skill reports which tasks are blocking and which upstream tasks must complete first.

Related Skills

Explore More Software Engineering Skills→
dixusdixus
10

decompose

Decompose concept documents into independent PRDs with acceptance criteria and dependencies.

Community
Advanced
pinoshepinoshe

github

Automate GitHub pull request workflows from branch creation to PR opening.

Community
Advanced
mikemylmikemyl
1

mutation-testing

Execute Stryker mutation testing on JavaScript and TypeScript codebases.

Community
Advanced
koala-man-64koala-man-64

audit-workitem

Decompose work items into small verifiable tasks for single-session coding agents.

Community
Intermediate