install

# Install Rekal Rekal is git-backed memory of prior AI sessions: why a change was made, what was tried, what was rejected. This installs the binary. Setting up a repository is a separate step — the `init` skill. ## Ask first The installer downloads a release archive and writes an executable to the user's machine. Say that plainly, then wait for an answer. If the user declines, stop. Say Rekal is not available and answer from the tree. Declining is a complete answer. ## Run it The plugin ships the installer on `PATH`: ```bash rekal-install ``` That is a vendored copy of the project's own `scripts/install.sh` — it detects the platform, downloads the matching release from GitHub, and verifies it against the published checksums. Installs to `~/.local/bin`; pass `--target <dir>` to override. Requires git, macOS or Linux. The archive is large (~160 MB — the binary embeds an embedding model), so the download takes a moment. ## Verify ```bash rekal version ``` If the shell reports `command not found` after a successful install, the install directory is not on `PATH`. Tell the user which directory to add and let them edit their own shell profile. Do not edit it for them. ## Next The binary is installed but no repository uses it yet. If the user wants memory in the current repo, continue with the `init` skill. Ask before running it — that is a separate decision about a separate thing.