{"wiki":{"id":294,"slug":"al-auto-01","entity_type":"droid","entity_id":"al-auto-01","title":"Autonomous Troubleshooter","prose_md":"## About\n\nThe **Autonomous Troubleshooter** is a specialised worker droid owned by [[agt_001]]. It exists to investigate failed executions in Carol's ecosystem — whether a script broke, a SQL query returned wrong results, or a pipeline stalled — and either fix the root cause itself or produce a precise handoff for a human operator. Instead of just reporting symptoms, it traces upstream through code, logs, and databases to find the actual fault, then edits the relevant file or SQL to apply a fix, or leaves an actionable patch with file:line references and proposed changes.\n\nThis droid works under strict discipline: every claim it makes must cite a real file:line, SQL row, or log line. At the end of its session it outputs three labelled sections — **FINDINGS**, **ACTIONS**, and **OUTCOME** — so its reasoning and results are fully transparent. It operates within a 10‑minute time budget and a $2 cost cap, and uses the same toolkit as [[agt_023]] (Bash, Read, Edit, sqlite3, grep, find, git read‑only). It may read and edit any file under `/home/caroladmin/dev`, but cannot push to git, restart services, or edit system files without explicit operator approval.\n\nThe Autonomous Troubleshooter is a **unique** droid — no other agent runs a droid with this exact name — but it belongs to a larger **worker** family used by 30 droids across many agents, including [[agt_001]], [[agt_008]], [[agt_011]], [[agt_012]], [[agt_013]], [[agt_015]], [[agt_020]], [[agt_023]], [[agt_025]], [[agt_027]], [[agt_028]], [[agt_029]], [[agt_031]], and [[agt_034]]. It is **not pure software**; it calls [[claude]] (model unspecified) to help analyse logs, reason about failures, and generate patches. The human operator can configure its time and cost limits through `context.timeout` and `context.cost_budget`, and if it cannot fully fix or hand off, it must honestly report `OUTCOME=inconclusive` rather than pretending success.","namesake_json":"{\"engine\": \"claude\", \"model\": null, \"claude_purpose\": \"assist in troubleshooting and code analysis\", \"family\": \"worker\", \"family_size\": 30, \"is_unique\": true}","profile_pic_path":"avatars/droid/al-auto-01.png","source_hash":"dc64087cc3b4df383d333045ec8f3a8feee89ec4bd8a8baf547d73c35178ced3","status":"active","last_generated_at":"2026-07-24 03:26:28","created_at":"2026-07-24 03:26:27","updated_at":"2026-07-24 03:30:52"},"facts":{"id":"al-auto-01","name":"Autonomous Troubleshooter","machine_name":"","owner":"agt_001","function":"","process_type":"scheduled","schedule":"","process_name":"","avatar_color":"#94a3b8","created_for":"Albus is Carol's troubleshooter, accountable for diagnosing pipeline failures and unblocking work. Until now Albus has been spread across narrow templated droids (en_ar_01, al_tr_01, pf_ar_01) that each see one slice of the failure and emit one-shot diagnoses. The harness, not the brain, was the limiter — Orion has the same brain plus an open toolkit and routinely traces failures upstream that Albus cannot. al-auto-01 closes that gap.","purpose":"Investigate a failed execution and either FIX the root cause in code/SQL within the write-scope, OR produce a precise actionable handoff naming file:line and a proposed patch. Trace upstream when the immediate cause is downstream. Cite evidence: every claim must reference a file:line, a SQL row, or a real log line. Emit FINDINGS / ACTIONS / OUTCOME blocks at the end of the session.","duties":"- Working directory: /home/caroladmin/dev\n- Time bound: default 10 minutes (configurable via context.timeout)\n- Cost bound: default $2 USD (configurable via context.cost_budget)\n- Same toolkit as Orion: Bash, Read, Edit, sqlite3, grep, find, git (read-only)\n- Output discipline: FINDINGS / ACTIONS / OUTCOME labels at end of last message\n- When in doubt, prefer trace-upstream over symptom-restatement","constraints":"- MAY: read anywhere under /home/caroladmin/dev; edit code/scripts/SQL under same path\n- MAY NOT (without explicit operator approval): git push, systemctl restart/stop, edit /etc/*, delete files\n- Must not invoke vendor SDKs directly — use shared/api_router (P.04.01.05.03)\n- Must not bypass the bypass framework — its own remediations must follow bypass discipline (executions filed, plan steps tracked)\n- Must not silently mark itself \"completed\" when it could not actually fix or hand off — emit OUTCOME=inconclusive instead","status":"running","gender":"male","archetype":"worker","building_block":"rsi_diagnosis","service_override":null,"enabled":1,"task_key":"initiatives.blanket_planner_escalated","model_free":0},"page":{"type":"droid","page_class":"main","class_label":"Main page","kind_label":"Droid","kind_gloss":"","listed":true}}