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.

ssn - Prompt - OpenSmartRoute

Name: ssn
Author: Barqly

Prompt file imported from Barqly/barqly-vault (.claude/commands/ssn.md). Copyright stays with the author.

Resume from Previous Session

Load context from the most recent session summary and continue work, optionally with a specific agent.

Instructions

Step 1: Handle Agent Selection (if provided)

Check if an agent was specified: {ARG}

If {ARG} is provided:

Switch to the {ARG} agent before loading context
Agent name shortcuts:
- sbe → sr-backend-engineer
- jbe → jr-backend-engineer
- sfe → sr-frontend-engineer
- jfe → jr-frontend-engineer
- sa or arch → system-architect
- po → product-owner
- zm or zen → zenmaster
- re → research-engineer
- devops or do → devops-engineer
- ux or uxd → ux-designer
- qa → qa-engineer
- Full agent names also work

If {ARG} is empty:

Continue with current/default agent

Step 2: Load Session Context

Find and read the most recent quick start file in handoff/ssc*.md or handoff/ssd*-quick.md
Also reference the corresponding detailed summary handoff/ssd*.md for full context
If multiple files exist, use the most recent one

Step 3: Present Understanding and Wait

Present understanding filtered through the active agent's perspective:

🔄 Session Resumed
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

🤖 Agent: [Current agent - either {ARG} or default]
📚 I've loaded the context from your previous session. Here's my understanding:

**What we were working on:**
[Brief summary of the main objective - from agent's perspective]

**Current state:**
[Where things stand - what's working, what's not]

**Key decisions/insights from last session:**
- [Important point 1]
- [Important point 2]
- [Important gotcha or constraint]

**Next steps we identified:**
1. [Step 1]
2. [Step 2]
3. [Step 3]

**Files in focus:**
- `[file]` - [why]
- `[file]` - [why]

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

✅ Ready to continue! Would you like me to:
- Proceed with the next step we identified?
- Address something specific first?
- Review anything from the previous session?

Waiting for your input before making any changes...

Usage

/ssn - Resume with current/default agent
/ssn sr-backend-engineer - Resume with backend engineer agent
/ssn fe - Resume with frontend engineer agent (using shortcut)

Important

Do NOT start coding immediately
Do NOT make file changes without confirmation
DO wait for user direction
DO be ready to adjust based on user's current priorities
The user might want to pivot or address something different first
If agent switch fails, continue with current agent but notify user
Agent switch (if any) happens BEFORE loading context to avoid waste

Install ssn 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.

# Resume from Previous Session Load context from the most recent session summary and continue work, optionally with a specific agent. ## Instructions ## Step 1: Handle Agent Selection (if provided) Check if an agent was specified: `{ARG}` If `{ARG}` is provided: - Switch to the `{ARG}` agent before loading context - Agent name shortcuts: - `sbe` → sr-backend-engineer - `jbe` → jr-backend-engineer - `sfe` → sr-frontend-engineer - `jfe` → jr-frontend-engineer - `sa` or `arch` → system-architect - `po` → product-owner - `zm` or `zen` → zenmaster - `re` → research-engineer - `devops` or `do` → devops-engineer - `ux` or `uxd` → ux-designer - `qa` → qa-engineer - Full agent names also work If `{ARG}` is empty: - Continue with current/default agent ## Step 2: Load Session Context 1. Find and read the most recent quick start file in `handoff/ssc*.md` or `handoff/ssd*-quick.md` 2. Also reference the corresponding detailed summary `handoff/ssd*.md` for full context 3. If multiple files exist, use the most recent one ## Step 3: Present Understanding and Wait Present understanding filtered through the active agent's perspective: ``` 🔄 Session Resumed ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 🤖 Agent: [Current agent - either {ARG} or default] 📚 I've loaded the context from your previous session. Here's my understanding: **What we were working on:** [Brief summary of the main objective - from agent's perspective] **Current state:** [Where things stand - what's working, what's not] **Key decisions/insights from last session:** - [Important point 1] - [Important point 2] - [Important gotcha or constraint] **Next steps we identified:** 1. [Step 1] 2. [Step 2] 3. [Step 3] **Files in focus:** - `[file]` - [why] - `[file]` - [why] ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ✅ Ready to continue! Would you like me to: - Proceed with the next step we identified? - Address something specific first? - Review anything from the previous session? Waiting for your input before making any changes... ``` ## Usage - `/ssn` - Resume with current/default agent - `/ssn sr-backend-engineer` - Resume with backend engineer agent - `/ssn fe` - Resume with frontend engineer agent (using shortcut) ## Important - Do NOT start coding immediately - Do NOT make file changes without confirmation - DO wait for user direction - DO be ready to adjust based on user's current priorities - The user might want to pivot or address something different first - If agent switch fails, continue with current agent but notify user - Agent switch (if any) happens BEFORE loading context to avoid waste

{ "prompt": "# Resume from Previous Session\n\nLoad context from the most recent session summary and continue work, optionally with a specific agent.\n\n## Instructions\n\n## Step 1: Handle Agent Selection (if provided)\n\nCheck if an agent was specified: `{ARG}`\n\nIf `{ARG}` is provided:\n- Switch to the `{ARG}` agent before loading context\n- Agent name shortcuts:\n - `sbe` → sr-backend-engineer\n - `jbe` → jr-backend-engineer \n - `sfe` → sr-frontend-engineer\n - `jfe` → jr-frontend-engineer\n - `sa` or `arch` → system-architect\n - `po` → product-owner\n - `zm` or `zen` → zenmaster\n - `re` → research-engineer\n - `devops` or `do` → devops-engineer\n - `ux` or `uxd` → ux-designer\n - `qa` → qa-engineer\n - Full agent names also work\n\nIf `{ARG}` is empty:\n- Continue with current/default agent\n\n## Step 2: Load Session Context\n\n1. Find and read the most recent quick start file in `handoff/ssc*.md` or `handoff/ssd*-quick.md`\n2. Also reference the corresponding detailed summary `handoff/ssd*.md` for full context\n3. If multiple files exist, use the most recent one\n\n## Step 3: Present Understanding and Wait\n\nPresent understanding filtered through the active agent's perspective:\n\n```\n🔄 Session Resumed\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n🤖 Agent: [Current agent - either {ARG} or default]\n📚 I've loaded the context from your previous session. Here's my understanding:\n\n**What we were working on:**\n[Brief summary of the main objective - from agent's perspective]\n\n**Current state:**\n[Where things stand - what's working, what's not]\n\n**Key decisions/insights from last session:**\n- [Important point 1]\n- [Important point 2]\n- [Important gotcha or constraint]\n\n**Next steps we identified:**\n1. [Step 1]\n2. [Step 2]\n3. [Step 3]\n\n**Files in focus:**\n- `[file]` - [why]\n- `[file]` - [why]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n✅ Ready to continue! Would you like me to:\n- Proceed with the next step we identified?\n- Address something specific first?\n- Review anything from the previous session?\n\nWaiting for your input before making any changes...\n```\n\n## Usage\n\n- `/ssn` - Resume with current/default agent\n- `/ssn sr-backend-engineer` - Resume with backend engineer agent\n- `/ssn fe` - Resume with frontend engineer agent (using shortcut)\n\n## Important\n\n- Do NOT start coding immediately\n- Do NOT make file changes without confirmation\n- DO wait for user direction\n- DO be ready to adjust based on user's current priorities\n- The user might want to pivot or address something different first\n- If agent switch fails, continue with current agent but notify user\n- Agent switch (if any) happens BEFORE loading context to avoid waste", "variables": [], "notes": "Load context from the most recent session summary and continue work, optionally with a specific agent.", "metadata": { "source": { "provider": "github-claude-commands", "repository": "https://github.com/Barqly/barqly-vault", "path": ".claude/commands/ssn.md", "ref": "1d9c89d1a05b906dbc574c5df57a68390f63dbde", "url": "https://github.com/Barqly/barqly-vault/blob/1d9c89d1a05b906dbc574c5df57a68390f63dbde/.claude/commands/ssn.md", "key": "Barqly/barqly-vault/.claude/commands/ssn.md" } } }

ssn

About

Resume from Previous Session

Instructions

Step 1: Handle Agent Selection (if provided)

Step 2: Load Session Context

Step 3: Present Understanding and Wait

Usage

Important

Install ssn 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