project

Manage long-running projects with a durable repo-based tasks.md tracker.

25|1|Updated Feb 10, 2026
One-click install
npx skills add https://github.com/wisdom-in-a-nutshell/agents --skill project-wisdom-in-a-nutshell
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: project
Source: https://github.com/wisdom-in-a-nutshell/agents/tree/main/skills-source/owned/project
Command: npx skills add https://github.com/wisdom-in-a-nutshell/agents --skill project-wisdom-in-a-nutshell

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? Long-running, multi-session work loses context between agent runs, leaving stale plans, forgotten decisions, and no clear resume point. This Skill keeps a single project tracker file in the repo as the durable source of truth so work can be created, resumed, replanned, and closed out cleanly. ## Core Features & Use Cases - Durable project tracker: Creates and maintains a tasks.md tracker with milestones, a live Current Batch execution board, backlog, decisions, blockers, and a progress log. - Resume and replan workflow: Detects missing, active, stale, or finished trackers and resumes from Current Batch or rebuilds the plan in place before continuing. - Subagent delegation conventions: Provides rules for splitting read-heavy exploration, external research, and bounded implementation work while keeping the tracker single-writer. - Safe closeout and archival: Bundles scripts/archive_project.py to atomically move the complete project directory into the tracker home's archive folder with validated preconditions and JSON output. - Use Case: Ask an agent to carry a multi-week refactor across sessions; it records every batch, decision, and validation result in the tracker, then archives the whole project directory when the Done When criteria pass. ## Quick Start Use the project skill to create a tracker for my migration project, plan the first batch of work, and keep it updated until the project is done and archived.

Frequently Asked Questions about project

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

FAQPage Schema
How do I resume a long-running agent project across sessions?▼

Keep a tasks.md tracker in the repo with a Current Batch section as the resume point. On each new session, read Current Batch, Milestones, Open Questions, and the Progress Log first, then continue execution from there instead of rescanning the repo.

Where should the project tracker file live in a repository?▼

Follow repo-local guidance first, such as STRUCTURE.md or AGENTS.md defining a tracker home like projects/<project>/tasks.md. If no guidance exists, use the default docs/projects/<project>/tasks.md, and archive under the same tracker home.

When should I use subagents for project work?▼

Delegate when the batch has independent questions answerable in parallel, preferring read-heavy exploration and external research first. Keep work local when the shared contract is still moving, tasks are tiny, or coordination cost exceeds the speedup.

How do I archive a completed project tracker safely?▼

Run the bundled archive_project.py script with --source and --destination pointing at the tracker home's archive directory. It validates the layout, moves the whole directory atomically, and reports source_removed: true on success.

Why does archive_project.py reject my archive destination?▼

The script enforces a strict layout: the destination must sit in an archive directory sibling to the active project, share the project name, and not already exist. It also rejects symlink sources and directories missing tasks.md.

When should a project tracker not be archived?▼

Skip archiving when completion is materially uncertain, validation failures are unresolved, or Open Questions / Blockers contains items that could change the deliverable. Record the concrete blocker in the tracker instead of leaving a ready-to-archive placeholder.