core

Enforces project-management conventions for markdown artefacts, status gates, and lifecycle commands.

Updated May 25, 2026
One-click install
npx skills add https://github.com/DATA-AI-XYZ/Tandem --skill core-data-ai-xyz
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: core
Source: https://github.com/DATA-AI-XYZ/Tandem/tree/main/skills/core
Command: npx skills add https://github.com/DATA-AI-XYZ/Tandem --skill core-data-ai-xyz

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When working with AI assistants on a project, decisions, plans, and test evidence get lost in chat scrollback, and nothing stops half-finished work from being marked done. This Skill enforces a governed PM operating kit where every artefact is a versioned markdown file and every status transition passes a mechanical gate. ## Core Features & Use Cases - Lifecycle command chain: Defines the canonical order of slash commands from OKRs through PRD, epic, story, testplan execution, and close-out, with a fork for single-story versus batched phase paths. - Mandatory artefact rules: Enforces a closed 9-value status enum, ISO 8601 frontmatter timestamps, Story-Testplan pairing, automatic bug filing on test failure, and ADR creation for non-obvious decisions. - DoR/DoD gates and hat protocol: Blocks stories from entering in-progress without a Definition of Ready check, blocks done without a Definition of Done check, and separates Founder, PM, Dev, and QA responsibilities across sessions. - Use Case: While implementing a story, the assistant creates the paired testplan in the same response, files a bug when a test case fails, records an ADR for a schema naming decision, and updates MONITOR.md when the story flips to done. ## Quick Start Ask the assistant to create a new story with its paired testplan under the project management folder and follow the Tandem PM conventions.

Frequently Asked Questions about core

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

FAQPage Schema
How do I enforce Definition of Ready and Definition of Done gates with Claude Code?

This Skill blocks status transitions until checklists pass: a story cannot enter in-progress without meeting the DoR checklist in SOP.md section 6, and cannot reach done without the DoD checklist in section 7. Missing items stop the workflow and prompt the user.

How do I keep project decisions from getting lost in AI chat history?

Record every non-obvious decision as an ADR markdown file in the decisions folder in the same response where the decision is made. ADRs are numbered sequentially and linked from the story's decisions array, keeping rationale versioned next to the code.

What status values are allowed for stories and bugs?

Exactly nine values form a closed enum: not-started, ready, in-progress, in-review, done, blocked, wontfix, duplicate, and archived. Invented values like open, shipped, or completed are rejected, and each status flip must set the matching ISO 8601 timestamp in the same edit.

Does the skill support different project folder layouts?

Yes, folder names resolve through a pm-paths script driven by .claude-pm-config.json, which supports full, flattened, or custom layouts. Skills reference canonical names like 32-Stories, and the configuration maps them to the project's actual directories.

What happens when a test case fails during story execution?

A bug artefact is filed automatically in the same response, before reporting in chat. The bug uses the ID format BUG-YYYYMMDD-NN, includes reproduction steps, an environment snapshot, a first analysis hypothesis, and a suggested fix direction.