Update to Goldsky

{
  "description": "Turbo pipeline operations reference — lifecycle commands (pause, resume, restart, delete), pipeline states, checkpoint behavior, streaming vs job-mode differences, CLI syntax for `inspect`/`logs`, TUI shortcuts, and error pattern lookup. Triggers on: 'how do I pause/restart/delete', 'will deleting lose my data', 'what does this error mean', 'inspect TUI shortcuts'. For interactive diagnosis of a broken pipeline, use /turbo-doctor.",
  "included_files": [
    {
      "relative_path": "data/error-patterns.json",
      "size_in_bytes": 10627
    },
    {
      "relative_path": "evals/trigger-eval.json",
      "size_in_bytes": 2715
    },
    {
      "relative_path": "scripts/analyze-logs.sh",
      "size_in_bytes": 5744
    }
  ],
  "name": "turbo-operations",
  "skill_md_contents": "---\nname: turbo-operations\ndescription: \"Turbo pipeline operations reference — lifecycle commands (pause, resume, restart, delete), pipeline states, checkpoint behavior, streaming vs job-mode differences, CLI syntax for `inspect`/`logs`, TUI shortcuts, and error pattern lookup. Triggers on: 'how do I pause/restart/delete', 'will deleting lose my data', 'what does this error mean', 'inspect TUI shortcuts'. For interactive diagnosis of a broken pipeline, use /turbo-doctor.\"\n---\n\n# Turbo Pipeline Operations\n\nLifecycle commands, monitoring, and error reference for running Turbo pipelines. This is a lookup reference — for interactive troubleshooting of a broken pipeline, use `/turbo-doctor`. For building new pipelines, use `/turbo-builder`.\n\n## Pipeline States\n\n| State     | Description                                  |\n| --------- | -------------------------------------------- |\n| running   | Pipeline is actively processing data         |\n| starting  | Pipeline is initializing                     |\n| paused    | Pipeline is paused (replicas set to 0)       |\n| stopped   | Pipeline is not running (manually stopped)   |\n| error     | Pipeline encountered an error                |\n| completed | Job-mode pipeline finished processing range  |\n\n## Streaming vs Job Mode Lifecycle\n\n| Operation     | Streaming Pipeline            | Job-Mode Pipeline (`job: true`)             |\n| ------------- | ----------------------------- | ------------------------------------------- |\n| **List**      | Shows as `running`/`paused`   | Shows as `running`/`completed`              |\n| **Pause**     | Supported                     | Not supported                               |\n| **Resume**    | Supported                     | Not supported                               |\n| **Restart**   | Supported                     | Not supported — use delete + apply          |\n| **Delete**    | Supported                     | Supported (auto-cleanup ~1hr after done)    |\n| **Apply**     | Updates in place              | Must delete first, then re-apply            |\n\n### Job-Mode Pipeline Lifecycle\n\nJob-mode pipelines (`job: true`) are one-time batch processes:\n\n1. **Start** — process data from `start_at` to `end_block`\n2. **Run** — process the bounded data range\n3. **Complete** — automatically stop when range is processed\n4. **Auto-cleanup** — ~1 hour after completion, automatically removed\n\nCannot pause, resume, or restart. Must delete before redeploying.\n\n## Lifecycle Commands\n\n### List Pipelines\n\n```bash\ngoldsky turbo list\n```\n\n### Pause a Pipeline\n\nTemporarily stop processing without deleting. Preserves all state for later resumption.\n\n```bash\ngoldsky turbo pause <pipeline-name>\n```\n\n### Resume a Pipeline\n\nRestore a paused pipeline. Can only resume a **paused** pipeline.\n\n```bash\ngoldsky turbo resume <pipeline-name>\n```\n\n### Restart a Pipeline\n\nTrigger a pod restart for a running or paused pipeline.\n\n```bash\ngoldsky turbo restart <pipeline-name>\n# To clear all checkpoints and reprocess from the beginning:\ngoldsky turbo restart <pipeline-name> --clear-state\n```\n\n### Delete a Pipeline\n\nPermanently remove a pipeline. All checkpoints are lost. Data already written to sinks is preserved.\n\n```bash\ngoldsky turbo delete <pipeline-name>\n```\n\n### Delete and Recreate (Fresh Start)\n\n```bash\ngoldsky turbo delete my-pipeline\ngoldsky turbo apply my-pipeline.yaml\n```\n\n## Checkpoint Behavior\n\n- Deleting a pipeline removes its checkpoints permanently\n- Recreating with the same name starts fresh (no checkpoint recovery)\n- To preserve checkpoints, use `apply` to update instead of delete/recreate\n- Checkpoint state is tied to source names — renaming a source resets its checkpoint\n- Checkpoint state is tied to pipeline names — renaming a pipeline resets all checkpoints\n\n## Monitoring Commands\n\n| Action                   | Command                                              |\n| ------------------------ | ---------------------------------------------------- |\n| List pipelines           | `goldsky turbo list`                                 |\n| View live data           | `goldsky turbo inspect <name> -p`                    |\n| Inspect specific node    | `goldsky turbo inspect <name> -n <node> -p`          |\n| View logs                | `goldsky turbo logs <name>`                          |\n| Follow logs              | `goldsky turbo logs <name> --follow`                 |\n| Logs with timestamps     | `goldsky turbo logs <name> --timestamps`             |\n| Last N lines             | `goldsky turbo logs <name> --tail N`                 |\n| Logs since N seconds ago | `goldsky turbo logs <name> --since N`                |\n\n## `goldsky turbo inspect` Flags\n\n| Flag                   | Short | Description                                                  |\n| ---------------------- | ----- | ------------------------------------------------------------ |\n| `--print`              | `-p`  | Print records to stdout                                      |\n| `--topology-node-keys` | `-n`  | Comma-separated node keys to filter (e.g. a transform name)  |\n| `--buffer-size`        | `-b`  | Max records to keep in buffer (default: 10000)               |\n\nAlways use `-p`. Always include it in every `inspect` command you suggest.\n\n### Log Analysis Script\n\nUse the helper script to quickly analyze pipeline logs:\n\n```bash\n./scripts/analyze-logs.sh <pipeline-name>\n./scripts/analyze-logs.sh <pipeline-name> --tail 100\n```\n\nThe script checks for common error patterns and reports findings with recommendations.\n\n## Common Error Patterns\n\n> **Detailed error patterns and solutions are in `data/error-patterns.json`.**\n\n| Error Pattern             | Likely Cause               | Fix                                            |\n| ------------------------- | -------------------------- | ---------------------------------------------- |\n| `connection refused`      | Database unreachable       | Check network/firewall settings                |\n| `authentication failed`   | Wrong credentials          | Update secret with correct credentials         |\n| `secret not found`        | Missing secret             | Create secret with `goldsky secret create`     |\n| `SQL syntax error`        | Invalid transform SQL      | Fix SQL in YAML and redeploy                   |\n| `duplicate key`           | Primary key collision      | Ensure unique primary key in transform          |\n| `script transform error`  | TypeScript runtime failure | Check script logic, null handling, return types |\n| `dynamic_table` error     | Backend connection issue   | Verify dynamic table secret/table exists        |\n| `WASM execution failed`   | Script crash in sandbox    | Debug script — check for undefined access       |\n| `handler timeout`         | External HTTP endpoint slow| Increase `timeout_ms` or fix handler endpoint   |\n\n### Script Transform Issues\n\n| Issue                        | Fix                                                    |\n| ---------------------------- | ------------------------------------------------------ |\n| `undefined` property access  | Add null checks: `input.field ?? ''`                   |\n| Wrong return type            | Ensure returned object matches `schema` exactly        |\n| Missing return fields        | All `schema` fields must be present in returned object |\n| `invoke is not a function`   | Ensure script defines `function invoke(data)`          |\n| BigInt errors                | Use `BigInt()` constructor, not direct number literals  |\n\n### Dynamic Table Issues\n\n| Issue                      | Fix                                                    |\n| -------------------------- | ------------------------------------------------------ |\n| Table not found            | Create the table in PostgreSQL before deploying         |\n| No matches from check      | Verify data exists in the backing table                 |\n| Stale data                 | For postgres backend, verify rows are actually there    |\n| Memory pressure            | Large in_memory tables → switch to postgres backend     |\n\n## Troubleshooting Quick Reference\n\n| Symptom                  | Likely Cause           | Quick Fix                                       |\n| ------------------------ | ---------------------- | ----------------------------------------------- |\n| No data flowing          | `start_at: latest`     | Wait for new data or use `earliest`             |\n| Auth failed              | Wrong credentials      | Update secret with correct password             |\n| Connection refused       | Network/firewall       | Check host, whitelist Goldsky IPs               |\n| Storage exceeded         | Neon free tier (512MB) | Upgrade plan or clear data                      |\n| SQL error                | Bad transform syntax   | Validate YAML first                             |\n| Pipeline not found       | Name mismatch          | Run `goldsky turbo list` to check names         |\n| Permission denied        | Role insufficient      | Verify Editor or Admin role in the project      |\n| `pipeline already exists`| Job-mode stale         | Delete first, then re-apply                     |\n| Cannot pause/resume job  | Job-mode limitation    | Job pipelines don't support pause/resume        |\n| Cannot restart job       | Job-mode limitation    | Delete + re-apply instead                       |\n| Can't connect to inspect | Pipeline not running   | Check status with `goldsky turbo list`          |\n| Logs are empty           | Pipeline just started  | Wait for data or check `start_at`               |\n| TUI disconnects          | Pipeline interrupted   | Auto-reconnects within 30 min; check status     |\n\n## Related\n\n- **`/turbo-doctor`** — Interactive diagnostic skill for pipeline issues\n- **`/turbo-builder`** — Build and deploy new pipelines\n- **`/turbo-pipelines`** — YAML configuration and architecture reference\n"
}