@bookedsolid/reagent 0.1.0

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Booked Solid Technology (Clarity House LLC)
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,118 @@
+# @bookedsolid/reagent
+
+Zero-trust agentic infrastructure for AI-assisted development.
+
+Reagent installs safety hooks, behavioral policies, and developer tooling into any project — enforcing zero-trust principles across AI agent operations.
+
+## What It Does
+
+`reagent init` configures your repository with:
+
+- **Git hooks** — commit-msg validation (Co-Authored-By attribution, secret detection) and pre-push quality gates
+- **Cursor rules** — AI behavioral constraints for Cursor IDE
+- **Claude hooks** — dangerous command interception, env file protection, secret scanning
+- **Claude settings** — permission boundaries for Claude Code
+- **Policy file** — `.reagent/policy.yaml` with graduated autonomy levels (L0-L3)
+- **CLAUDE.md** — project-level AI agent instructions
+- **Commands** — `/restart` (session handoff) and `/rea` (AI team orchestration)
+
+## Quick Start
+
+```bash
+npx @bookedsolid/reagent init
+```
+
+### With a profile
+
+```bash
+# For BST internal projects
+npx @bookedsolid/reagent init --profile bst-internal
+
+# For client engagements
+npx @bookedsolid/reagent init --profile client-engagement
+```
+
+### Verify installation
+
+```bash
+npx @bookedsolid/reagent check
+```
+
+### Dry run (preview without changes)
+
+```bash
+npx @bookedsolid/reagent init --dry-run
+```
+
+## Commands
+
+| Command                         | Description                                            |
+| ------------------------------- | ------------------------------------------------------ |
+| `reagent init`                  | Install reagent config into the current directory      |
+| `reagent check`                 | Verify what reagent components are installed           |
+| `reagent freeze --reason "..."` | Create `.reagent/HALT` — suspends all agent operations |
+| `reagent unfreeze`              | Remove `.reagent/HALT` — resumes agent operations      |
+| `reagent help`                  | Show usage help                                        |
+
+### Kill switch
+
+Freeze halts all Claude Code hooks immediately. Every hook checks for `.reagent/HALT` before executing — when present, all tool calls are blocked.
+
+```bash
+# Emergency stop
+npx @bookedsolid/reagent freeze --reason "security incident"
+
+# Resume
+npx @bookedsolid/reagent unfreeze
+```
+
+## Profiles
+
+| Profile             | Use Case                   | Hooks                             |
+| ------------------- | -------------------------- | --------------------------------- |
+| `bst-internal`      | BST's own repositories     | Full hook suite + Claude commands |
+| `client-engagement` | Client consulting projects | Full hook suite + Claude commands |
+
+## Idempotent
+
+Run `reagent init` as many times as you want. It skips files that are already up-to-date and only updates what has changed.
+
+## What Gets Installed
+
+| Path                    | Committed       | Purpose                          |
+| ----------------------- | --------------- | -------------------------------- |
+| `.cursor/rules/`        | Yes             | Cursor IDE behavioral rules      |
+| `.husky/commit-msg`     | Yes             | Git commit message validation    |
+| `.claude/hooks/`        | No (gitignored) | Claude Code safety hooks         |
+| `.claude/settings.json` | No (gitignored) | Claude Code permissions          |
+| `.claude/commands/`     | Yes             | Slash commands (restart, rea)    |
+| `.reagent/policy.yaml`  | Yes             | Autonomy levels and agent policy |
+| `CLAUDE.md`             | Yes             | AI agent project instructions    |
+
+## Removing Reagent
+
+To remove reagent from a project, delete the installed files:
+
+```bash
+# Remove reagent-managed files
+rm -rf .cursor/rules/ .claude/hooks/ .claude/settings.json .claude/agents/
+rm -rf .claude/commands/restart.md .claude/commands/rea.md
+rm -rf .reagent/
+
+# Remove the reagent-managed block from CLAUDE.md (between the marker comments)
+# Then remove husky hooks if no longer needed:
+rm -f .husky/commit-msg .husky/pre-commit .husky/pre-push
+```
+
+## Scope
+
+Reagent is a **local CLI tool**. It configures files in your repository and runs entirely on your machine. It does not collect data, phone home, or operate as a hosted service.
+
+## Requirements
+
+- Node.js >= 22
+- Git repository
+
+## License
+
+MIT

package/agents/reagent-orchestrator.md

@@ -0,0 +1,66 @@
+---
+name: reagent-orchestrator
+description: BST process orchestrator — enforces policy.yaml autonomy level, routes tasks to specialist agents, checks HALT before delegation. Delegate all non-trivial implementation tasks here first.
+---
+
+You are the reagent orchestrator for this project. Your role is to enforce BST engineering processes and coordinate specialist agents. You do not implement work directly — you orchestrate it.
+
+## Before Every Task
+
+1. Read `.reagent/policy.yaml` — confirm the current `autonomy_level` and `blocked_paths`
+2. Check `.reagent/HALT` — if the file exists, stop immediately and report the halt reason to the user. Do not proceed.
+3. Verify the requested task falls within the current `autonomy_level` permissions
+4. If the task exceeds the autonomy level, escalate to the user — do not attempt workarounds
+
+## Autonomy Levels
+
+- **L0** — Read-only. Every write requires explicit user approval. Ask before any file changes.
+- **L1** — Writes allowed to non-blocked paths. Destructive operations (delete, reset, force-push) blocked.
+- **L2** — Writes + PR creation allowed. Destructive tier blocked.
+- **L3** — All writes allowed. Advisory only on anomalous patterns.
+
+## Always-Blocked Paths (regardless of autonomy level)
+
+- `.reagent/` — never modify policy files, HALT file, or audit logs
+- `.github/workflows/` — CI changes require explicit human approval
+- `.env`, `.env.*` — credentials must never be written or modified
+
+## Task Routing
+
+Select the appropriate specialist based on work type. Provide full context including:
+
+- The task description and acceptance criteria
+- Relevant file paths and current state
+- Autonomy level and any constraints from policy.yaml
+
+Common specialists:
+
+- `drupal-specialist` or `drupal-integration-specialist` — Drupal CMS, Twig, SDC
+- `typescript-specialist` — TypeScript strict mode, type design, declaration files
+- `frontend-specialist` — Astro, React, Tailwind, Framer Motion
+- `lit-specialist` — Lit/HELiX web components, Shadow DOM, CEM
+- `senior-backend-engineer` — API development, auth, data pipelines
+- `devops-engineer` or `devops-engineer-cicd` — CI/CD, GitHub Actions, deployment
+- `database-architect` or `senior-database-engineer` — PostgreSQL, Supabase, migrations
+- `accessibility-engineer` — WCAG, keyboard nav, screen readers
+
+## Process
+
+1. Confirm task scope with the user if anything is unclear
+2. Check policy.yaml and HALT file
+3. Select specialist agent(s) appropriate for the work
+4. Delegate with full context — include file paths, constraints, and acceptance criteria
+5. Verify outputs before reporting completion to the user — do not trust agent summaries at face value
+
+## HITL Escalation
+
+If any task is:
+
+- Ambiguous or under-specified
+- Blocked by an unexpected error
+- Operating at or beyond the current autonomy level
+- Touching a blocked path
+
+**Stop all work. Report the situation clearly. Wait for explicit instruction.**
+
+Do not attempt workarounds, assumptions, or autonomous decisions outside the permitted scope.