AI Agent Hub
Back to plugins
dsh-odoo-sdd preview

dsh-odoo-sdd

Model Inference Updated 2026.09.16

Run the following command in DeepSeek Harness:

dsh plugin install fhidalgodev/dsh-odoo-sdd

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install fhidalgodev/dsh-odoo-sdd inside DeepSeek Harness to install this plugin; the project is open-sourced at https://github.com/fhidalgodev/dsh-odoo-sdd

About this plugin

Odoo development rarely benefits from a single structured loop: specs, code, and verification live in separate tools, and AI assistants have no way to install a module on a real instance, read the traceback, and retry. dsh-odoo-sdd turns DeepSeek Harness into a specification-driven pipeline — spec, architecture, code, verify — where every phase persists to disk behind a fail-closed gate. It never spawns Docker; it points at a dev or staging instance you already run, speaks plain JSON-RPC to install modules, and feeds ir.logging or server tracebacks straight back to the model. Three consecutive failures force a root-cause diagnosis, five iterations force BLOCKED, and a failed verification is always persisted as FAILED, never reframed as a win.

Beyond module development, the plugin offers a functional path for configuration and data work on a live instance: changes are planned in batches, each approval is bound to content hashes, environment mismatches are refused outright, and production writes require a declared backup reference plus a separate sign-off. Imports go through Odoo's own importer rather than hand-written parsers. A static analysis layer (structure and ACL coherence, sudo, auth=none, t-raw scans) produces file:line findings before anything is installed, and documentation generation (OCA readme fragments, Apps index page, changelog) works standalone with no spec or instance required.

Ideal for developers and implementation consultants who pair AI with Odoo module work, bug fixes, or greenfield setups (companies, taxes, users, master data). Requirements: DSH 0.1.2-rc.1 or later, Node 20+, an HTTP-reachable Odoo instance, and a disposable test database. Playwright is optional and only activates the UI-verification tier.

Screenshots

Use Cases

  • Develop an Odoo module spec-to-verify in a closed feedback loop against a real instance
  • Configure a company, taxes, and master data on staging with hash-bound batch approvals
  • Run static security scans and generate OCA docs without booting an instance

Best For

  • Developers building or fixing Odoo modules with AI assistance
  • Consultants implementing Odoo company, user, and master-data setups for clients
  • Teams that must verify AI-generated code against a live instance rather than a paper review