work-issue

Execute a tracker issue end-to-end with TDD until acceptance criteria are met.

5|Updated Mar 30, 2026

Search Tags:#automation #tdd #software-engineering #issue-tracking #pr-workflow #blocker-report #loop-execution

Ready to run

npx skills add https://github.com/kenlck/skills --skill work-issue-kenlck

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: work-issue
Source: https://github.com/kenlck/skills/tree/main/ken-loop/skills/work-issue
Command: npx skills add https://github.com/kenlck/skills --skill work-issue-kenlck

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Headless executor for one loop-ready issue: reads the issue, implements it with TDD against its acceptance criteria, and ends in exactly one of two exits, DELIVERED (branch + PR) or BLOCKED (findings comment + needs-human).

Core Features & Use Cases

Deterministic, non-interactive execution of a single issue per run.
Branch-and-PR workflow: creates a loop/issue-<n> branch, reports progress, and validates acceptance criteria.
Clear exit paths with evidence: DELIVERED (branch + PR) or BLOCKED (findings comment + needs-human).

Quick Start

Invoke this skill on a specific tracker issue to execute it end-to-end with TDD until acceptance criteria are met.

Frequently Asked Questions about work-issue

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

FAQPage Schema

How do I automate issue execution end-to-end using TDD against acceptance criteria?▼

Automating issue execution with TDD requires a non-interactive executor that reads a single tracker issue, implements the code via test-driven development, and validates against explicit acceptance criteria. The process ends in either a DELIVERED or BLOCKED state.

Can I use headless execution to generate a pull request for a single tracker issue?▼

Headless execution can generate a pull request for a single tracker issue by creating a deterministic loop/issue-<n> branch, implementing the changes, and validating the acceptance criteria before reporting progress and exiting with a DELIVERED state.

How does a non-interactive workflow report blockers during automated issue execution?▼

Non-interactive workflows report blockers during automated issue execution by terminating the run with a BLOCKED exit state, logging evidence in a findings comment, and flagging the issue as needs-human to ensure the blocker is reviewed.

Does automated issue execution support multiple issues per run?▼

Automated issue execution does not support multiple issues per run, as the design strictly enforces deterministic, non-interactive execution of a single tracker issue per invocation to ensure accurate validation against its specific acceptance criteria.

What is the best way to enforce deterministic branching for TDD issue automation?▼

The best way to enforce deterministic branching for TDD issue automation is to map each single tracker issue to a specific loop/issue-<n> branch, ensuring non-interactive execution tracks progress and validates acceptance criteria accurately before a PR is created.

Why does automated TDD issue execution require explicit acceptance criteria?▼

Automated TDD issue execution requires explicit acceptance criteria because the headless executor validates the implementation against these predefined conditions to determine whether the run should exit as DELIVERED or BLOCKED with a findings comment.