Specstride core built on it: Hermes host · Pins the core 0.1.0 release tarballs by sha256.
Hermes host
Specstride inside Hermes — from a chat, the TUI or Telegram, derive, dry-run, launch, supervise, stop, resume and curate Spec Kit features through the same pinned engine as the other hosts. Push notices to Telegram come with it, and a kanban board of the runs is an install option.
Built from a clone: nothing is on npm yet. Linux is verified, macOS is not. The pinned engine has no reverse verb yet, so the installer skips the reverse skill.
It is the same engine as the Claude Code plugin and the dsh host. The package adds no
product logic and copies no engine source: every fact it shows is a line of the CLI,
reached only through the shim specstride-hermes, and the engine arrives as
sha256-pinned tarballs under vendor/core/ (core-pin.json). The skills are rendered
from @specstride/host-kit, so the three hosts cannot drift; Python exists only as
thin plugin glue that forwards to Node.
The installer installs, reports and removes. It never touches the default Hermes
profile: --profile specstride enables the plugin in a dedicated profile, and --user
writes files and prints the enable command instead of running it. The doctor prints,
after specstride doctor, one <severity> <check> — <fix> line per finding — eight
checks, no model, within 2 s on the reference machine.
Two things the other hosts do not have. The watcher unit sends one Telegram line per
run state change or attention item — a failed delivery is one log line and never gates
a run. And install --with-kanban adds the projector module for a kanban board: one
root card per run, one stage card per phase, one blocked card per attention item. A
board write made by a person is undone by the restore guard; a human-edited title or
body is kept and gets one stale comment instead.
Skills
specstride-run- Run Specstride on the finished features of the repository — derive, judge, validate, render, launch and supervise to the digest. Invoking the skill is the consent to launch;
dry-runstops after the plan. specstride-status- Show the state of a run on this Hermes host from the precomputed status line, without starting anything.
specstride-curate- Reconcile the Spec Kit corpus of the repository for the Hermes host — backup, checks, retargeting rules, analyze refresh.
Install
Hermes
Build the Hermes host from a clone, install it with the engine's CLI from the tarball it pins, then put the skills and the plugin into a dedicated Hermes profile. Nothing is on npm yet.
Needs
- Linux. macOS is not verified.
- Node 22+ and git
- Hermes 0.22.2 (the pinned hermes-agent), with a profile
specstride(hermes profile create specstrideif absent) - The engine CLI on PATH or
SPECSTRIDE_BINset - A Spec Kit feature in a git repository
In one command — runs the steps below, in order (read the script)
curl -fsSL https://specstride.ai/install.sh | bash -s -- hermes
Or step by step
git clone https://github.com/mairp/specstride-hermes cd specstride-hermes # resolves the pinned engine tarballs; nothing compiles npm install # renders the skills and the shim, runs every build gate npm run build # specstride-host-hermes and the specstride CLI this repository pins npm install -g . vendor/core/specstride-0.1.0.tgz # skills and plugin into the dedicated profile; run it again → already installed, unchanged specstride-host-hermes install --profile specstride # eight host.* findings, one line each, no model specstride-host-hermes doctor
Then, in a Hermes session on your repository (hermes -p specstride), say "run the specstride-run skill in dry-run on the feature". It plans and validates and launches nothing. Drop "in dry-run" to run the feature for real. Add --with-kanban to the install for the kanban board.
Commands last run as shown: 2026-11-14
RepositoryStar mairp/specstride-hermes on GitHub
Useful? A star on mairp/specstride-hermes and on the engine, mairp/specstride-core, helps others find Specstride.