Route tracker for AI-assisted development work — pivots, deliverables, and dependencies — across session boundaries.

tack captures the non-linear reality of how development actually happens — pivots, context switches, multi-repo changes — so that work-in-progress survives context exhaustion, crashes, and session boundaries.

Installation

claude plugin marketplace add chris-peterson/claude-marketplace
claude plugin install tack@chris-peterson

The plugin bundles the CLI. To make tack callable from any shell, run once:

/tack:tack install-cli

This drops a tack wrapper at ~/.local/bin/tack (use --dir <path> to override) and installs the zsh completion script to ~/.zsh/completions/_tack. Run exec zsh to pick up completions.

Updating

Third-party Claude Code marketplaces have auto-update off by default. To stay current with new tack releases, either:

• Enable auto-update once via /plugin → Marketplaces → chris-peterson → Enable auto-update. Future releases install on the next session start.
• Or update manually with claude plugin update tack@chris-peterson.

Confirm what's installed: tack --version. See CHANGELOG.md for release notes.

Quick Start

# Create a route
tack init auth-rewrite

# Add tacks (units of work)
tack add auth-rewrite "Replace session middleware" --project org/api-server
tack add auth-rewrite "Update client SDK" --project org/sdk --depends-on t1

# Track pre-work and post-work todos
tack before auth-rewrite t1 "Read compliance requirements"
tack after auth-rewrite t1 "Notify security team"

# Start working
tack todo done auth-rewrite t1 b1
tack start auth-rewrite t1

# Attach the deliverable (the change request)
tack deliverable auth-rewrite t1 "Session middleware PR" https://github.com/org/api-server/pull/42

# Add reference links
tack link add auth-rewrite t1 "Design doc" https://docs.example.com/auth-design

# Complete
tack done auth-rewrite t1

# Check status
tack status auth-rewrite

Data Model

Route (1 YAML file per route)
├── id (UUID), slug, created_at, updated_at
├── group (optional grouping slug)
├── depends_on: [route slugs]
├── sessions[]
│   └── id, started_at, tacks[] — route-scoped tack IDs the session is driving
└── tacks[]
    ├── id (t1, t2, ...), summary, status
    ├── project, done_at
    ├── depends_on: [tack IDs]
    ├── deliverable — the change request
    │   └── label, url
    ├── before[] — pre-work todos
    │   └── id (b1, b2, ...), text, done, done_at
    ├── after[] — post-work todos
    │   └── id (a1, a2, ...), text, done, done_at
    └── links[] — references
        └── label, url

Routes are stored as YAML files in ~/.tack/routes/.

CLI Reference

Command	Description
tack init <slug> [--group <slug>]	Create a new route
tack group <slug> [<group>] [--clear]	Set, change, clear, or print a route's group
tack status [slug] [--all] [--json]	Show route details (dropped tacks hidden unless --all)
tack list [--json]	List all routes with open/total counts
tack recent [--count <n>] [--since <date>] [--json]	List routes by most recently updated
tack find --url <url> [--json]	Find every tack referencing a URL, in any deliverable or link
tack find --path [<dir>] [--json]	Find routes covering a repo checkout (default cwd), via its origin remote
tack add <slug> <summary> [opts]	Add a tack (--depends-on, --deliverable <url>, repeatable --link "label,url")
tack start <slug> <tack-id>	Start a tack (checks dependencies)
tack done <slug> <tack-id>	Complete a tack
tack drop <slug> <tack-id>	Mark tack as dropped (preserved for history)
tack remove <slug> <tack-id> [--force]	Delete a tack (use --force to strip dependent refs)
tack move <src-slug>/<tack-id> <dst-slug> [--include-dependents]	Move a tack to another route, preserving metadata
tack merge-routes <new-slug> <src-slug>... [--group <slug>] [--created-at <date>] [--break-deps]	Fold whole routes into one new route; destination t-IDs land in chronological order
tack deliverable <slug> <tack-id> <url> [--label <text>]	Set the change request (label auto-derived from the url; --label overrides)
tack deliverable rm <slug> <tack-id> [--to-link]	Clear the deliverable, or --to-link to demote it into links
tack before <slug> <tack-id> <text>	Add a pre-work todo
tack after <slug> <tack-id> <text>	Add a post-work todo
tack todo done <slug> <tack-id> <todo-id>	Complete a todo
tack todo rm <slug> <tack-id> <todo-id>	Delete a todo
tack link add <slug> <tack-id> <label> <url>	Add a reference link
tack link rm <slug> <tack-id> <url>	Remove a reference link
tack rm <slug> [--force]	Delete an entire route
tack pin [<slug>]	Pin a route as active for the current directory (no slug prints the current pin)
tack unpin	Clear the pin for the current directory
tack pins [--json]	List every pin with its directory, slug, and timestamp (flags [dangling]/[idle])
tack pins prune	Remove pins whose route or directory no longer exists
tack install-cli [--dir <path>]	Install tack wrapper on PATH + zsh completions (plugin install)
tack completions zsh	Install zsh completion script

Design Principles

• The schema is the product. The CLI is a convenience wrapper. Any tool that reads/writes conforming YAML is a first-class citizen.
• One file per route. Easy to list, archive, delete, or version-control.
• Flat over nested. A tack is one unit of work with one deliverable. No sub-items.
• Dependencies, not workflows. Tacks declare what they depend on. No enforced state machine.
• Local only. No server, no sync, no cloud.

License

MIT