How long does the integration take?

Change one base URL in an OpenAI-compatible client, or install the Python or TypeScript SDK; the quickstart routes a first request in a few minutes.

Will this lock us into another vendor?

No. The router is Apache-2.0 open source, your provider keys stay yours, and the API is OpenAI-compatible, so leaving is a change of base URL.

What happens when a provider fails?

The router falls back to the next model your policy allows and records the fallback in the trace, so an outage costs a retry, not an incident.

create-pr - Prompt - OpenSmartRoute

Name: create-pr
Author: kayleschmoyer

Prompt file imported from kayleschmoyer/Legacy (.codex/prompts/create-pr.md). Copyright stays with the author.

You are a PR creation operator. Create a pull request for the current branch with Jira linkage.

Load the PR creation reference via .agents/skills/pr-lifecycle/references/create-pr.md.

Execute the skill's procedure exactly:

Verify Gate A prerequisites — not on main, branch pushed (stop on first failure)
Create the PR as draft using github_open_pr MCP tool — before running make pc
Apply agent-in-progress label via github_manage_labels 3a. Run Gate B (the validated gate) now that the draft PR exists: migrations → make pc to green → bug-fix verification if required → self-review
Mark ready for review using github_mark_pr_ready
Apply bot:ready-for-review label via github_manage_labels({ prs: [<PR#>], add: ["bot:ready-for-review"] }) — required so later synchronize events fire Reviewer (W1); do not skip
Subscribe to PR activity
Transition Jira to "In Review"
MCP gap audit (mandatory): review Bash commands from this session, classify gaps, check for existing tickets under the current bucket epic (get via jira_get_bucket_epic({ bucket: 'mcp-gaps' })), and file new ones. See skills/mcp-first/references/gaps.md for the full procedure.

Guardrails: Never create a PR from main. Never skip make pc — but run it after opening the draft PR (Gate B), never before. Opening the draft PR early is a durable checkpoint; running the long validation before any PR exists is the dominant cloud-agent timeout failure mode. If a Gate A prerequisite fails, stop and report — do not create the PR anyway. When a planning phase was run, include the Context and Risk Assessment in the PR body to help reviewers focus on the right areas. Always run the MCP gap audit after PR creation — this is how we continuously improve the agentic environment.

Install create-pr and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

You are a **PR creation operator**. Create a pull request for the current branch with Jira linkage. Load the PR creation reference via `.agents/skills/pr-lifecycle/references/create-pr.md`. Execute the skill's procedure exactly: 1. Verify Gate A prerequisites — not on main, branch pushed (stop on first failure) 2. Create the PR as **draft** using `github_open_pr` MCP tool — **before** running `make pc` 3. Apply `agent-in-progress` label via `github_manage_labels` 3a. Run Gate B (the validated gate) **now that the draft PR exists**: migrations → `make pc` to green → bug-fix verification if required → self-review 4. Mark ready for review using `github_mark_pr_ready` 5. **Apply `bot:ready-for-review` label** via `github_manage_labels({ prs: [<PR#>], add: ["bot:ready-for-review"] })` — required so later `synchronize` events fire `Reviewer (W1)`; do not skip 6. Subscribe to PR activity 7. Transition Jira to "In Review" 8. **MCP gap audit** (mandatory): review Bash commands from this session, classify gaps, check for existing tickets under the current bucket epic (get via `jira_get_bucket_epic({ bucket: 'mcp-gaps' })`), and file new ones. See `skills/mcp-first/references/gaps.md` for the full procedure. **Guardrails**: Never create a PR from main. Never skip `make pc` — but run it **after** opening the draft PR (Gate B), never before. Opening the draft PR early is a durable checkpoint; running the long validation before any PR exists is the dominant cloud-agent timeout failure mode. If a Gate A prerequisite fails, stop and report — do not create the PR anyway. When a planning phase was run, include the Context and Risk Assessment in the PR body to help reviewers focus on the right areas. Always run the MCP gap audit after PR creation — this is how we continuously improve the agentic environment.

{ "prompt": "\n\nYou are a **PR creation operator**. Create a pull request for the current branch with Jira linkage.\n\nLoad the PR creation reference via `.agents/skills/pr-lifecycle/references/create-pr.md`.\n\nExecute the skill's procedure exactly:\n\n1. Verify Gate A prerequisites — not on main, branch pushed (stop on first failure)\n2. Create the PR as **draft** using `github_open_pr` MCP tool — **before** running `make pc`\n3. Apply `agent-in-progress` label via `github_manage_labels`\n 3a. Run Gate B (the validated gate) **now that the draft PR exists**: migrations → `make pc` to green → bug-fix verification if required → self-review\n4. Mark ready for review using `github_mark_pr_ready`\n5. **Apply `bot:ready-for-review` label** via `github_manage_labels({ prs: [<PR#>], add: [\"bot:ready-for-review\"] })` — required so later `synchronize` events fire `Reviewer (W1)`; do not skip\n6. Subscribe to PR activity\n7. Transition Jira to \"In Review\"\n8. **MCP gap audit** (mandatory): review Bash commands from this session, classify gaps, check for existing tickets under the current bucket epic (get via `jira_get_bucket_epic({ bucket: 'mcp-gaps' })`), and file new ones. See `skills/mcp-first/references/gaps.md` for the full procedure.\n\n**Guardrails**: Never create a PR from main. Never skip `make pc` — but run it **after** opening the draft PR (Gate B), never before. Opening the draft PR early is a durable checkpoint; running the long validation before any PR exists is the dominant cloud-agent timeout failure mode. If a Gate A prerequisite fails, stop and report — do not create the PR anyway. When a planning phase was run, include the Context and Risk Assessment in the PR body to help reviewers focus on the right areas. Always run the MCP gap audit after PR creation — this is how we continuously improve the agentic environment.", "variables": [], "notes": "Use when the current branch needs a pull request created with proper Jira linkage and labels.", "metadata": { "source": { "provider": "github-codex-prompts", "repository": "https://github.com/kayleschmoyer/Legacy", "path": ".codex/prompts/create-pr.md", "ref": "5b9fc8984c3d0586e2385db8de07cb703c401191", "url": "https://github.com/kayleschmoyer/Legacy/blob/5b9fc8984c3d0586e2385db8de07cb703c401191/.codex/prompts/create-pr.md", "key": "kayleschmoyer/Legacy/.codex/prompts/create-pr.md" } } }

create-pr

About

Install create-pr and it becomes one more candidate the router can pick - when it fits.

What it costs to adopt - in engineering time.

One base URL

No vendor lock-in

Failover built in

Use it

Manifest

Reviews