Intent CLI
Intent CLI keeps the goal, the latest verified checkpoint and decisions in the project so a new agent session can resume
- 266 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Project hooks run `itt` commands on agent events and the install downloads and runs a script.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/intent-cliDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Project hooks run `itt` commands on agent events and the install downloads and runs a script.
This is third-party code. Review the repository files before installing.
What it does
Intent CLI is a semantic history layer next to your code. The `itt init` command creates an `.intent` folder with plain JSON that both the agent and you can read. Recording is deliberately light: the agent saves a checkpoint only when something is verified, and a quiet turn writes nothing. Resuming is complete: one command gives the next session the goal, the verified state, the boundary, the next step and the rules that still apply. The history has three object kinds. Intent describes a goal and why it matters, Snap records a verified checkpoint with a boundary, blocker and next step, Decision keeps a durable rule for future goals. Everything is append-only and behaves like Git: `itt log --graph`, `itt diff`, `itt switch -c`, `itt merge`. Agents in separate git worktrees share one history, each on its own branch. For Claude Code there is a skill or plugin and project hooks. The IntHub web app is optional, and without it data stays on your machine.
Who it is for. Developers who run long tasks with Claude Code, Codex or Cursor and are tired of re-explaining to the agent where they left off.
Good fit when
- Tomorrow you continue yesterday task in a new session and the agent needs the goal, verified state and next step.
- A rule such as payments go only through PaymentIntent has to attach to new goals automatically.
- Several agents work in separate worktrees and their checkpoints must be merged.
- You need a history of reasons and completion boundaries, not only line changes.
Not a fit when
- A repository `AGENTS.md` or `CLAUDE.md` with rules is enough: Intent stores work state, not instructions.
- You need shared team planning with review in a tracker: issues and PRs cover that.
- You have no Python 3.9 or newer and cannot run the install script.
Example request
Resume this project through Intent: read the goal, the latest verified checkpoint and the standing decisions and name the next step.Limitations
The project is marked beta. Install uses `install.sh` for macOS and Linux or `install.ps1` for Windows, and `pipx install intent-history` and `uv tool install intent-history` also work, Python 3.9 or newer is required. The skill installs through Node.js with `npx skills add`. Project hooks need to be trusted once: through `/hooks` in Claude Code and in Codex as well. The agent records a checkpoint only after verification, so the history reflects verified work and not every turn. Data stays local until you choose to share it through IntHub.
How to disable. Remove the package with `pipx uninstall intent-history`, the skill and hooks with agent tools and the `.intent` folder from the project.
FAQ
How do I install Intent CLI?
With `curl -fsSL https://raw.githubusercontent.com/dozybot001/Intent/main/scripts/install.sh | bash` on macOS and Linux or `pipx install intent-history`. Then `itt init` in the project.
How do I install the skill in Claude Code?
With `npx skills add dozybot001/Intent -g -a claude-code -y` or the plugin: `/plugin marketplace add dozybot001/Intent` and `/plugin install intent-cli@intent`.
How do I resume in a new session?
Tell the agent: Resume this project through Intent. It reads the goal, verified state and rules.
Where is the history stored?
In the `.intent` folder inside the project as plain JSON. Git is optional.
How is it different from agent memory?
Agent memory is opaque inside one tool, while Intent sits in the project as open JSON that any agent and person can read.
Discussion
No questions or reviews yet.
Related
CLI toolsEncephalon stores project decisions and conventions as JSON inside Git so coding agents can search them between tasks.
A skill that turns a vague request into a typed IntentSpec: it looks answers up in your repo and asks the one question only you can answer
A portable .agent folder with memory, skills and protocols that plugs into many coding agents and keeps its knowledge when you switch tools
Shared infrastructure for Claude Code, Codex, Antigravity CLI and OpenCode: a task lifecycle, skills and sandboxes.
The ee CLI gives agents durable local memory: rules, decisions and anti-patterns with scores and provenance
Local memory for AI agents in plain Markdown files: it keeps only what you approve and serves it through MCP
README in short
The README explains the idea, quick install, a manual example, comparison with AGENTS.md and agent memory, and a table of supported hosts. MIT license, the PyPI package is called intent-history.