Skip to content

Open Mercato Skills - Setup Guide

Source: open-mercato/skills via skills.sh - 79.0K combined installs across 63 indexed listings; first seen Oct 9, 2026 (extended-query probe) GitHub: open-mercato/skills (231 stars, MIT; pushed Oct 5, 2026; skills live under skills/) Category: Enterprise ERP / Ecommerce Engineering Quality Tier: 🟡 Beta - company-backed (Open Mercato), MIT, active Oct 2026; Snyk Warn on sampled skills - see Security

Open Mercato Skills extracts the PR pipeline behind the Open Mercato ERP: inside that project the workflow produced ~800k lines of code with zero hand-written lines, 1700+ merged PRs, 4000 unit tests, 730 integration tests, and weekly releases with 100+ contributors. The repo ships 41 agent skills that run the full loop - plan, implement, review, QA gate, merge - stripped of everything product-specific so any team with a GitHub repo can run it with any coding agent.

Every skill reads one committed config file, .ai/agentic.config.json, written once per repo by om-setup-agent-pipeline, which inspects the default branch, validation scripts, and labels, then generates SDLC.md. The naming convention is the contract: the om-auto-* prefix means autonomous and non-interactive (safe to run on a schedule or in CI), while every skill without it is interactive and hands control back.

Installation

# Install all of the skills (any of 22+ coding agents, via skills.sh)
npx skills add open-mercato/skills --skill '*'

# Update later, from the project directory (-g for global installs)
npx skills update -p

Then, once per repository, run the configurator inside your agent:

/om-setup-agent-pipeline

It inspects your repo (default branch, validation scripts, GitHub labels), asks a few questions, writes .ai/agentic.config.json, and generates SDLC.md - the ticket-flow doc every other skill reads. After updating skills, /om-apply-upgrade-notes applies UPGRADE_NOTES.md migrations while preserving local edits. Drop --skill '*' to cherry-pick interactively.

What It Provides

Skill Installs Use For
om-apply-upgrade-notes 2,773 Post-upgrade migrator: applies UPGRADE_NOTES.md while preserving local edits
om-auto-continue-pr 2,763 Resumes an in-progress PR from its first unchecked step to completion
om-auto-create-pr 2,748 Task brief end-to-end: plan, worktree, commits, validation gate, labeled PR
om-auto-continue-pr-loop 2,738 Resumes loop runs from HANDOFF.md with per-step commits and checkpoints
om-auto-create-pr-loop 2,730 Long spec implementations: run folder, per-step commits, checkpoint verification
om-auto-fix-issue 2,707 The issue-to-PR entry point: classifies bug vs feature, then routes
om-auto-review-pr 2,603 Reviews a PR in an isolated worktree; autofix loop until merge-ready
om-approve-merge-pr 2,541 Approves and squash-merges a PR by number, with the QA-gate guard
om-auto-fix-pr 2,461 Drives one PR to merge-ready: review autofix, CI stabilization, UI QA
om-auto-implement-spec 2,377 Implements an existing spec, then review loop and UI screenshots on the PR
om-auto-manage-issues 2,365 Issue hygiene: labels, clarification comments, spec-coverage checks
om-auto-qa-pr 2,325 QAs a PR's UI in a real browser: screenshots and a pass/fail report, no merge
om-auto-update-changelog 2,216 Drafts a CHANGELOG entry per merged PR and ships it as a docs PR
om-code-review 2,123 The review checklist behind om-auto-review-pr: correctness, security, contracts
om-setup-agent-pipeline 2,059 One-per-repo configurator: config, SDLC.md, cross-skill coverage check
om-integration-tests 2,036 Integration/E2E tests against the running app: real locators, no hardcoded IDs
om-fix 2,031 Implements the minimal fix with regression tests; runs the validation gate
om-auto-write-spec 2,015 Turns a brief or feature issue into a finished spec PR with mockups
om-open-pr 2,011 Shared PR opener: commit, push, unified body, SDLC labels, chain markers
om-prepare-issue 2,004 Files one well-formed issue: dedupe, linked spec, SDLC labels on creation

The remaining 43 indexed listings range from 1 to 1,969 installs.

Why This Matters for Hermes Agents

This collection is a battle-tested PR pipeline an agent can drive end to end: hand it a brief, a spec, or an issue number and it plans, implements in an isolated worktree, runs your validation gate, self-reviews, QAs the UI in a real browser when needed, and finishes with a ready, fully labeled PR. The discipline is built in, which matters when agents work unattended - claim locks so concurrent agents back off instead of colliding, chain markers so skills hand off without duplicating work, and a QA gate that refuses to merge a needs-qa PR without qa-approved even when checks are green. It is stack-agnostic: the base branch, validation commands, and label taxonomy come from one committed config file, so the same pipeline runs on a Rust, Go, or TypeScript repo. The skills are designed to be extended rather than forked - repo-local overrides at .ai/skills//SKILL.md take precedence, and custom tracker or browser providers plug in through committed descriptor files. One caveat to plan for: a few skills drive a real browser, and the README notes skills.sh validation may flag those as Medium or High risk - read them before you run them.

Usage

You say What happens
"/om-setup-agent-pipeline" Configures the repo: writes the config, generates SDLC.md, verifies cross-skill coverage.
"/om-auto-create-pr \"add rate limiting to the login endpoint\"" Plans, implements phase by phase in a worktree, runs validation, self-reviews, opens a labeled PR.
"/om-auto-fix-issue 123" Classifies the issue, then routes: bug to the autofix chain, feature to spec-then-implement.
"/om-auto-qa-pr 123" Boots the app, drives the browser provider, and posts screenshots with a pass/fail report.
"/om-approve-merge-pr 123" Approves and squash-merges; refused when needs-qa lacks qa-approved or a blocking label is set.
"/om-auto-update-changelog" Drafts a CHANGELOG entry for every PR since the last release and ships it as a docs PR.
"/om-discover --mode client \"Benefits portal for SMB clients\"" Builds a product-brief.md from real material, with evidence-tagged claims and owned decisions.

Verification

Confirm the skills are installed and the repo config exists, then fetch a skill's full source before running it:

# List installed skills
npx skills list | grep om-

# Confirm the repo config after the setup step
cat .ai/agentic.config.json

# Review a skill's full source before install
curl -sL https://raw.githubusercontent.com/open-mercato/skills/main/skills/om-apply-upgrade-notes/SKILL.md

Security

skills.sh verdicts for sampled skills (verified Oct 10, 2026) - re-check the security pages on skills.sh before production use:

Skill Gen Agent Trust Hub Socket Snyk
om-apply-upgrade-notes Pass Pass Warn
om-auto-continue-pr Pass Pass Warn
om-auto-create-pr Pass Pass Warn

Integration with CorpusIQ

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects

Limitations

  • Snyk Warn on all three sampled skills: treat the pipeline as privileged automation and re-check the security pages on skills.sh before production use.
  • Three skills (om-prepare-test-env, om-integration-tests, om-auto-qa-pr) drive a real browser; the README notes skills.sh validation may flag them Medium or High risk - read them before running.
  • The pipeline expects a configured repo: without .ai/agentic.config.json (written by om-setup-agent-pipeline), skills run setup first or lack the context they need.
  • The README frames the collection as 41 agent skills while skills.sh indexes 63 listings; treat the repo catalog as the canonical skill set.
  • New publisher repo first seen Oct 9, 2026: external validation of the collection is still limited even though the workflow shipped a real product.

  • Updates flow through npx skills update plus /om-apply-upgrade-notes; skip the upgrade skill and UPGRADE_NOTES.md migrations are not applied.

  • Snapshot data, verified Oct 10, 2026: 79,035 combined installs across 63 indexed listings; 231 GitHub stars; MIT; last pushed Oct 5, 2026. Counts drift over time.

Pro Tips

  • Run /om-setup-agent-pipeline once per repo before anything else; every skill reads the config it writes.
  • Prefer the autonomous entry points: /om-auto-create-pr for a brief, /om-auto-fix-issue for an issue number.
  • Extend without forking: drop a repo-local override at .ai/skills//SKILL.md; local rules win but can never relax safety rules.
  • After upgrading, run /om-apply-upgrade-notes so UPGRADE_NOTES.md migrations are applied without clobbering your local edits.
  • Keep --skill '*' for the first install so cross-skill references and chain markers resolve; the pipeline composes.

← Skills Catalog | Skills Marketplace → Powered by CorpusIQ

AI answers on your data. See where every number comes from.

Connect your tools and ask in ChatGPT, Claude, or Perplexity. Enterprise-grade answers in minutes. From $29.95/month.

Start free, no credit card
Contact Terms Privacy Pricing