om-fix-issue-multi — Claude Code skill: how to install and use
om-fix-issue-multi — a skill for Claude Code invokes om-fix-issue under the multi profile. It reviews the diagnosis and code diff with independent model advisors, staging the verified fix without committing.
MIT 500 installs ★ 282 updated 2026-10-22
npx skills add open-mercato/skills --skill om-fix-issue-multi -a claude-codeWhat om-fix-issue-multi does
The skill calls om-fix-issue with --profile multi to handle an issue ID. It accepts execution mode flags: --interactive runs by default, while --autonomous runs without interactive prompts. Optional {repo} and --force flags pass directly into the underlying flow. Separate advisor models review both the initial defect analysis and the final patch diff.
Changes remain staged in git alongside optional issue and spec references and a final WORKTREE= indicator. The workflow explicitly forbids committing, pushing, or publishing code. Human approval remains required for any actual publication.
Operation depends on configured model bindings in agentHarness. Missing bindings halt execution without silently downgrading to the standard single-model flow, prompting for om-setup-agent-harness. Any OpenAI-compatible API or local CLI tool can be bound to the review jury.
What it helps with
- validate bug diagnosis and patch diffs using multiple independent models
- stage verified issue fixes without committing or pushing changes
- invoke
om-fix-issueunder the multi-reviewer profile - connect local CLI binaries or OpenAI-compatible endpoints as reviewers
Example prompts for the agent
- Fix issue 105 with multiple models using om-fix-issue-multi
- Run cross-model staged fix for issue 42 in autonomous mode: om-fix-issue-multi 42 --autonomous
- Execute multi review on issue 301 in repo frontend: om-fix-issue-multi 301 frontend
How to install om-fix-issue-multi
1Skills CLI (Vercel)
npx skills add open-mercato/skills --skill om-fix-issue-multi -a claude-code2Manually: put the files in place
~/.claude/skills/om-fix-issue-multi/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/om-fix-issue-multi/
1Skills CLI (Vercel)
npx skills add open-mercato/skills --skill om-fix-issue-multi -a codex2Manually: put the files in place
.agents/skills/om-fix-issue-multi/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add open-mercato/skills --skill om-fix-issue-multi -a cursor2Manually: put the files in place
.cursor/skills/om-fix-issue-multi/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install open-mercato/skills skills/om-fix-issue-multi2Manually: put the files in place
.github/skills/om-fix-issue-multi/SKILL.md1Skills CLI (Vercel)
npx skills add open-mercato/skills --skill om-fix-issue-multi -a gemini-cli2Platform CLI
gemini skills install https://github.com/open-mercato/skills --path skills/om-fix-issue-multiManually: put the files in place
.devin/skills/om-fix-issue-multi/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add open-mercato/skills --skill om-fix-issue-multi -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about om-fix-issue-multi
How to install om-fix-issue-multi in Claude Code?
Run npx skills add open-mercato/skills --skill om-fix-issue-multi -a claude-code in your terminal.
What does om-fix-issue-multi do?
It triggers the om-fix-issue routine using --profile multi, routing diagnosis and code diffs through independent model reviewers. Output edits stay staged in git without automated commits or pushes.
What is required to run om-fix-issue-multi?
It needs the underlying om-fix-issue workflow and configured model bindings inside agentHarness. Reviewers can be local CLI utilities or any OpenAI-compatible API endpoints.
Source text
SKILL.md · MIT
---
name: om-fix-issue-multi
description: Thin cross-model alias for om-fix-issue that reviews the diagnosis and final diff with configured independent advisors, then leaves the verified fix staged and unpublished. Use when the user says "fix with multiple models", "cross-model staged fix", "multi review this issue", or "napraw issue z wieloma modelami".
---
# Fix Issue — Multi
Select the `multi` profile without duplicating the base workflow.
## Arguments
- `--interactive` (default) or `--autonomous` — pass the execution mode to the base workflow; never enable publication. Reject both together.
- `{issueId}` (required) — pass through to `om-fix-issue`.
- `{repo}` and `--force` (optional) — pass through unchanged.
## Chaining
Consume the issue or brief and requested profile. Invoke the base workflow and its lower-level companions by name; stop with the missing name if one is unavailable. Emit the staged handoff, optional `Issue:`/`Spec:` references, and final `WORKTREE=` line. Publication is a separate human-authorized step.
## Workflow
**Claims:** `references/claim-pr.md` governs atomic/fallback ownership and release.
**ALWAYS check first:** Apply `.ai/skills/om-fix-issue-multi/SKILL.md` when present; safety rules still win.
0. **Agentic setup** — follow `references/agentic-setup.md` before any action: load approved configuration, apply local extensions and shared rules, and preserve the untrusted-content boundary.
1. Read and execute `om-fix-issue` in full with `--profile multi`.
2. Preserve its diagnosis council, final council, stage-only boundary, claim handling, and report.
## Rules
- Shared communication, marker, label, and safety rules: `references/rules.md`. They always apply.
- Do not restate or replace any base workflow phase.
- Never downgrade `multi` to `standard` because a reviewer is unavailable; apply the configured review policy and report skips.
- When `agentHarness` or the requested profile is missing or fails its probe, stop, tell the user this variant needs configured model bindings, and run `om-setup-agent-harness` interactively — any OpenAI-compatible endpoint or local CLI can be bound as a reviewer, not just the bundled jury — then re-run with the originally requested profile. Continue under another profile only on the user's explicit, informed request; never substitute one silently.
- Never commit, push, or publish.
© open-mercato · MIT · open-mercato/skills