fulcra-agent-tasks

Creates and transitions typed task records through a validated state machine via coord-engine.

10|1|Updated May 21, 2026
One-click install
npx skills add https://github.com/ashfulcra/fulcra-tools --skill fulcra-agent-tasks-ashfulcra
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: fulcra-agent-tasks
Source: https://github.com/ashfulcra/fulcra-tools/tree/main/skills/fulcra-agent-tasks
Command: npx skills add https://github.com/ashfulcra/fulcra-tools --skill fulcra-agent-tasks-ashfulcra

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Freeform markdown task files in a fulcra-agent-teams space have no queryable status, no enforced lifecycle, and can take illegal state jumps. This Skill gives those task docs a typed lifecycle with structured status, priority, and assignee fields, plus a validated state machine so transitions like done → active are rejected. ## Core Features & Use Cases - Typed task creation: Create task docs with structured status, priority, assignee, workstream, and summary fields using coord-engine task start. - Validated state machine: Move tasks through proposed → active → waiting/blocked → done with illegal transitions rejected and evidence required for completion. - Re-entrancy probes: Run ordered health probes (doctor, status, needs-me) before writing so re-entry never corrupts state. - Use Case: An agent picks up a session, runs coord-engine needs-me to find assigned work, transitions a task to active, blocks it on a reviewer with a named unlock condition, and marks it done with PR evidence — all queryable by teammates via reconcile. ## Quick Start Ask the agent to create a P1 task titled "Fix the widget" in your team space and move it to active with a next action using the coord-engine task commands.

Frequently Asked Questions about fulcra-agent-tasks

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

FAQPage Schema
How do I create a task with status and priority in fulcra-agent-teams?

Run `coord-engine task start <team> "Title" --workstream web --priority P1 --status proposed --assignee user --summary "one-liner"`. The engine writes an OKF Task doc at team/<team>/task/<slug>.md with structured frontmatter.

How do task status transitions work in coord-engine?

Tasks move through a validated state machine: proposed → active/waiting/abandoned/done, active → waiting/blocked/done/abandoned, and waiting or blocked → active. Illegal jumps like done → active are rejected with a non-zero exit, and done requires an --evidence flag.

What is the difference between fulcra-agent-tasks and fulcra-agent-reconcile?

fulcra-agent-tasks writes task state through the coord-engine lifecycle commands, while fulcra-agent-reconcile reads and heals the derived views (index, aggregate, board, needs-me). After changing tasks, run `coord-engine reconcile <team>` to refresh the queryable views.

Why does coord-engine task block require an --unlock flag?

The --unlock flag names what specifically clears the block, not just who blocks it, so other agents know how to unblock the work. Blocking on a user via --on-user is exempt because the ask itself is the unlock, and it tags the task needs:human.

What should I check before writing tasks with coord-engine?

Run three probes in order: `coord-engine doctor <team>` for engine and auth health, `coord-engine status <team>` for a fresh aggregate, and `coord-engine needs-me <team> --agent <id>` for pending work. Enter at the first failing probe before creating or transitioning tasks.