[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77JW3M1FSRTHENTNKZ0N9WS":3,"skill-same:01M77JW3M1FSRTHENTNKZ0N9WS":50},{"skillNo":4,"name":5,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":13,"repositoryHtmlUrl":26,"repositoryHomepageUrl":6,"ownerHtmlUrl":27,"ownerBlogUrl":6,"ownerDescription":28,"updatedAt":29,"install":30,"resources":43},"01M77JW3M1FSRTHENTNKZ0N9WS","aionui-troubleshooting",null,"finewood2008\u002Fcentaurai-station-solo","finewood2008","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F212194470?v=4","01M75HGRZJ31P9W3T90VHFS1X8","Finewood","01M77JTDCYSHEZ7JPRS5B0CR5W",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002Fresources\u002Fbutler\u002Fskills\u002Faionui-troubleshooting\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: aionui-troubleshooting\ndescription: Diagnose a running CentaurAI installation — locate and inspect conversations (including stuck\u002Frunning ones), read aioncore logs, check LLM provider health, list scheduled cron jobs and their last run status, inspect teams and member state, and check MCP server health. Use when the user reports that CentaurAI is misbehaving: a conversation is stuck or errored, an LLM\u002Fprovider call is failing, a scheduled task did not run, an MCP server has no tools, a team member is hung, or they just ask \"what's wrong with CentaurAI\" \u002F \"排查一下 aionui\". Engine-agnostic — works the same for claude \u002F aionrs \u002F gemini \u002F openclaw conversations.\n---\n\n> **⚠️ Platform note — read before running any command.** The shell snippets in this skill are written for **macOS \u002F Linux** (bash\u002Fzsh). Always check which OS you are on first. On **Windows** do **not** run them verbatim — the underlying tool\u002FCLI commands are usually cross-platform, but the surrounding shell syntax is not. Translate it to PowerShell before running:\n>\n> | bash (macOS \u002F Linux)              | PowerShell (Windows)                                                                |\n> | --------------------------------- | ----------------------------------------------------------------------------------- |\n> | `a && b`                          | run as two steps, or `a; if ($?) { b }`                                             |\n> | `cat \u003C\u003C'EOF' \\| tool …` (heredoc) | write the text to a temp file, then pipe\u002Fpass that file to the tool                 |\n> | `VAR=$(cmd)` … `$VAR`             | `$VAR = cmd` … `$VAR`                                                               |\n> | `cmd > \u002Fdev\u002Fnull`                 | `cmd > $null`                                                                       |\n> | `… \\| grep PAT`                   | `… \\| Select-String PAT`                                                            |\n> | `… \\| jq …`                       | `… \\| ConvertFrom-Json`, then read the fields                                       |\n> | `python3 x.py`                    | `python x.py` (or `py x.py`)                                                        |\n> | `~\u002Fdir`, `\u002Ftmp`                   | `$env:USERPROFILE\\dir`, `$env:TEMP`                                                 |\n> | `cp` \u002F `mkdir -p` \u002F `rm -rf`      | `Copy-Item` \u002F `New-Item -ItemType Directory -Force` \u002F `Remove-Item -Recurse -Force` |\n>\n> If a command has no obvious Windows equivalent, prefer the built-in file\u002FHTTP tools over raw shell.\n\n# CentaurAI Troubleshooting\n\nDiagnose a running CentaurAI installation by reading its **project-level** data:\nthe aioncore REST API, the unified SQLite store, and the aioncore log files.\n\nThis is **engine-agnostic**. CentaurAI runs conversations on several backends\n(`acp`\u002Fclaude, `aionrs`, `gemini`), but troubleshooting goes\nthrough CentaurAI's own data — the `conversations` API, the unified `messages`\ntable, provider health, crons, teams, MCP — so the same checks work no matter\nwhich engine a conversation uses. Do **not** reach into engine-specific\ntranscript files (e.g. `~\u002F.claude\u002Fprojects\u002F*.jsonl` or\n`~\u002F.aionui\u002Faionrs-sessions\u002F*.json`); they are implementation details of one\nbackend and are already covered by the unified `messages` table.\n\n## How it works\n\nCentaurAI is front\u002Fback separated: the Electron UI talks to a local `aioncore`\nbackend. The backend's REST port is **dynamic** (aioncore launches with\n`--port 0`), so the first step is always discovery. The helper script discovers\neverything from the running process and wraps every read.\n\n## Setup\n\nOne self-contained helper — no external dependencies, read-only everywhere.\n\n```bash\npython3 scripts\u002Faion_diag.py discover\n```\n\n`discover` finds the live backend and prints what every other command relies on:\n\n```json\n{\n  \"pid\": 86716,\n  \"base_url\": \"http:\u002F\u002F127.0.0.1:58188\",\n  \"port\": 58188,\n  \"log_dir\": \"\u002FUsers\u002Fyou\u002FLibrary\u002FLogs\u002FCentaurAI\",\n  \"data_dir\": \"\u002FUsers\u002Fyou\u002F.aionui\",\n  \"version\": \"2.1.18\",\n  \"db_path\": \"\u002FUsers\u002Fyou\u002F.aionui\u002Faionui-backend.db\"\n}\n```\n\nHow discovery works (and why it's robust):\n\n- Finds the aioncore process started with `--data-dir` (the long-lived backend,\n  not the short-lived `mcp-guide-stdio` \u002F `mcp-team-stdio` helper subprocesses).\n- Reads `--log-dir`, `--data-dir`, `--app-version` straight from its argv — so\n  if the user changed the log directory, **we follow it**, never hardcode it.\n- The REST port is NOT in argv (`--port 0`), so the script probes every port the\n  process listens on and keeps the one that answers `\u002Fhealth` with `status:ok`.\n\nIf `discover` exits with an error \u002F code 3, CentaurAI is **not running**. Tell the\nuser to launch it — do not guess a port.\n\n> The script redacts secrets (`api_key`, tokens, …) in all output. Provider\n> records contain plaintext API keys; never print raw provider JSON yourself —\n> go through the helper.\n\n## Golden rule: start wide, then drill in\n\nFor \"something is wrong with CentaurAI\" with no specifics, run `overview` first —\nit's a one-shot snapshot across health, providers, MCP, crons, and running\nconversations. Then drill into whatever it flags.\n\n```bash\npython3 scripts\u002Faion_diag.py overview\n```\n\n---\n\n## Commands by symptom\n\nAll commands take the discovered backend automatically. Output is JSON unless\nnoted.\n\n### \"A conversation is stuck \u002F errored \u002F behaving wrong\"\n\n```bash\npython3 scripts\u002Faion_diag.py conversations [--limit N]   # list: id, name, engine type, status\npython3 scripts\u002Faion_diag.py conversation \u003Cid>           # status + runtime + recent errors + stuck hint\npython3 scripts\u002Faion_diag.py messages \u003Cid> [--limit N] [--errors]   # message history from SQLite\n```\n\n- `conversation \u003Cid>` is the workhorse. It returns the live `runtime` block\n  (`state`, `task_status`, `is_processing`, `turn_id`), the 5 most recent\n  **error** messages, and a `stuck_hint` when `state=running` +\n  `is_processing=true`.\n- **Stuck detection is comparative, not absolute.** A single `running` snapshot\n  is normal — that may just be the active turn. To confirm a hang, run\n  `conversation \u003Cid>` a few times seconds apart: if `turn_id` and runtime never\n  change while no new messages arrive, it's stuck. Cross-check with\n  `logs --conv \u003Cid>`.\n- `messages \u003Cid> --errors` pulls just the failed messages\u002Ftool-calls for that\n  conversation from the unified `messages` table (engine-agnostic).\n\n### \"An LLM \u002F provider call is failing\"\n\n```bash\npython3 scripts\u002Faion_diag.py providers\n```\n\nLists every configured provider with its `model_health` (`status`, `latency`,\n`last_check`) and an `unhealthy_models` summary. A provider whose models show\nnon-`healthy` status, huge `latency`, or a stale `last_check` is the suspect.\nThen confirm with the log (filter by the provider's base_url or id):\n\n```bash\npython3 scripts\u002Faion_diag.py logs --errors --lines 100\n```\n\n> `model_health` only holds the **most recent** check per model — there's no\n> historical error log in it. For the actual failure cause (timeout, 401, 429,\n> bad base_url), the aioncore log is the source of truth.\n\n### \"A scheduled task didn't run\"\n\n```bash\npython3 scripts\u002Faion_diag.py crons\n```\n\nThere is **no REST API for crons** — this reads the `cron_jobs` table from the\nSQLite store directly (read-only). It surfaces a `failing` list (jobs whose\n`last_status` is `error` or `missed`) plus every job's `schedule_*`,\n`last_status`, `last_error`, `next_run_at`, `last_run_at`, `run_count`,\n`retry_count`. Check `enabled`, compare `next_run_at` to now, and read\n`last_error` for failed jobs.\n\n### \"An MCP server has no tools \u002F isn't working\"\n\n```bash\npython3 scripts\u002Faion_diag.py mcp\n```\n\nLists MCP servers with `enabled`, `transport`, and `tool_count`. It flags any\nserver that is **enabled but exposes 0 tools** — the classic \"failed to start\"\nsignature (bad command, missing binary, crashed on launch). Then check the log\naround startup. A `disabled` server with 0 tools is fine — the user turned it off.\n\n### \"A team \u002F team member is hung\"\n\n```bash\npython3 scripts\u002Faion_diag.py teams\n```\n\nLists teams and members; for each member it resolves the linked\n`conversation_id` to its live `conv_state`. A member stuck in `running` (while\nothers are `idle`) is the one to drill into with `conversation \u003Cmember-conv-id>`.\n\n### \"Is the backend even alive? What version?\"\n\n```bash\npython3 scripts\u002Faion_diag.py health      # GET \u002Fhealth: status + core version + build_time\npython3 scripts\u002Faion_diag.py discover    # also shows app version, port, dirs\n```\n\n### Reading logs\n\n```bash\npython3 scripts\u002Faion_diag.py logs [--lines N] [--errors] [--conv \u003Cid>]\n```\n\n- Tails the latest `*.aioncore.log` in the discovered `log_dir` (NDJSON: one\n  JSON object per line — timestamp, level, target, HTTP status\u002Fpath).\n- `--errors` keeps only ERROR\u002FWARN\u002Ferror\u002Fpanic lines.\n- `--conv \u003Cid>` keeps only lines mentioning that conversation id — the fastest\n  way to see one conversation's request\u002Fresponse trace.\n- Need an arbitrary endpoint not wrapped above? `python3 scripts\u002Faion_diag.py\nget \u002Fapi\u002F\u003Cpath>` does a raw (redacted) GET.\n\n> Known log noise: `\"No onPostToolUseHook found for tool use ID: ...\"` WARNs fire\n> on nearly every tool call and are benign SDK-level chatter — don't treat them\n> as the fault.\n\n---\n\n## Data source map\n\n| Concern                           | Source                                  | Access                  |\n| --------------------------------- | --------------------------------------- | ----------------------- |\n| Backend alive \u002F version           | `GET \u002Fhealth`                           | REST                    |\n| Conversation list + runtime state | `GET \u002Fapi\u002Fconversations[\u002F{id}]`         | REST                    |\n| Conversation messages \u002F errors    | `messages` table (by `conversation_id`) | SQLite (read-only)      |\n| LLM provider health               | `GET \u002Fapi\u002Fproviders` → `model_health`   | REST (api_key redacted) |\n| Scheduled jobs                    | `cron_jobs` table                       | SQLite (no REST API)    |\n| Teams + members                   | `GET \u002Fapi\u002Fteams`                        | REST                    |\n| MCP servers                       | `GET \u002Fapi\u002Fmcp\u002Fservers`                  | REST                    |\n| Logs                              | `*.aioncore.log` in `--log-dir`         | File tail               |\n\n`db_path` and `log_dir` are always taken from `discover` (process argv), never\nhardcoded, so they track whatever the user configured.\n\n## Verification \u002F safety notes\n\n- All SQLite access is opened **read-only** (`mode=ro`) — diagnosis never mutates\n  the live store.\n- All output is run through secret redaction. If you ever need to show a provider\n  record, use the helper's `providers` command, not a raw `get`.\n- Confirm \"stuck\" only after repeated snapshots — one `running` reading is not a\n  fault.\n- This skill is **read-only diagnosis**. To _change_ configuration (create\u002Fedit\n  assistants, add MCP servers, attach skills), use the separate\n  `aionui-config` skill.\n\n## Not yet covered\n\n- No streaming\u002Flive log follow — only tail-on-demand.\n- No historical provider-error timeline (only the latest `model_health` check);\n  reconstruct history from the logs.\n- Repairing a broken conversation (vs. diagnosing it) is out of scope.\n","> **⚠️ Platform note — read before running any command.** The shell snippets in this skill are written for **macOS \u002F Linux** (bash\u002Fzsh). Always check which OS you are on first. On **Windows** do **not** run them verbatim — the underlying tool\u002FCLI commands are usually cross-platform, but the surrounding shell syntax is not. Translate it to PowerShell before running:\n>\n> | bash (macOS \u002F Linux)              | PowerShell (Windows)                                                                |\n> | --------------------------------- | ----------------------------------------------------------------------------------- |\n> | `a && b`                          | run as two steps, or `a; if ($?) { b }`                                             |\n> | `cat \u003C\u003C'EOF' \\| tool …` (heredoc) | write the text to a temp file, then pipe\u002Fpass that file to the tool                 |\n> | `VAR=$(cmd)` … `$VAR`             | `$VAR = cmd` … `$VAR`                                                               |\n> | `cmd > \u002Fdev\u002Fnull`                 | `cmd > $null`                                                                       |\n> | `… \\| grep PAT`                   | `… \\| Select-String PAT`                                                            |\n> | `… \\| jq …`                       | `… \\| ConvertFrom-Json`, then read the fields                                       |\n> | `python3 x.py`                    | `python x.py` (or `py x.py`)                                                        |\n> | `~\u002Fdir`, `\u002Ftmp`                   | `$env:USERPROFILE\\dir`, `$env:TEMP`                                                 |\n> | `cp` \u002F `mkdir -p` \u002F `rm -rf`      | `Copy-Item` \u002F `New-Item -ItemType Directory -Force` \u002F `Remove-Item -Recurse -Force` |\n>\n> If a command has no obvious Windows equivalent, prefer the built-in file\u002FHTTP tools over raw shell.\n\n# CentaurAI Troubleshooting\n\nDiagnose a running CentaurAI installation by reading its **project-level** data:\nthe aioncore REST API, the unified SQLite store, and the aioncore log files.\n\nThis is **engine-agnostic**. CentaurAI runs conversations on several backends\n(`acp`\u002Fclaude, `aionrs`, `gemini`), but troubleshooting goes\nthrough CentaurAI's own data — the `conversations` API, the unified `messages`\ntable, provider health, crons, teams, MCP — so the same checks work no matter\nwhich engine a conversation uses. Do **not** reach into engine-specific\ntranscript files (e.g. `~\u002F.claude\u002Fprojects\u002F*.jsonl` or\n`~\u002F.aionui\u002Faionrs-sessions\u002F*.json`); they are implementation details of one\nbackend and are already covered by the unified `messages` table.\n\n## How it works\n\nCentaurAI is front\u002Fback separated: the Electron UI talks to a local `aioncore`\nbackend. The backend's REST port is **dynamic** (aioncore launches with\n`--port 0`), so the first step is always discovery. The helper script discovers\neverything from the running process and wraps every read.\n\n## Setup\n\nOne self-contained helper — no external dependencies, read-only everywhere.\n\n```bash\npython3 scripts\u002Faion_diag.py discover\n```\n\n`discover` finds the live backend and prints what every other command relies on:\n\n```json\n{\n  \"pid\": 86716,\n  \"base_url\": \"http:\u002F\u002F127.0.0.1:58188\",\n  \"port\": 58188,\n  \"log_dir\": \"\u002FUsers\u002Fyou\u002FLibrary\u002FLogs\u002FCentaurAI\",\n  \"data_dir\": \"\u002FUsers\u002Fyou\u002F.aionui\",\n  \"version\": \"2.1.18\",\n  \"db_path\": \"\u002FUsers\u002Fyou\u002F.aionui\u002Faionui-backend.db\"\n}\n```\n\nHow discovery works (and why it's robust):\n\n- Finds the aioncore process started with `--data-dir` (the long-lived backend,\n  not the short-lived `mcp-guide-stdio` \u002F `mcp-team-stdio` helper subprocesses).\n- Reads `--log-dir`, `--data-dir`, `--app-version` straight from its argv — so\n  if the user changed the log directory, **we follow it**, never hardcode it.\n- The REST port is NOT in argv (`--port 0`), so the script probes every port the\n  process listens on and keeps the one that answers `\u002Fhealth` with `status:ok`.\n\nIf `discover` exits with an error \u002F code 3, CentaurAI is **not running**. Tell the\nuser to launch it — do not guess a port.\n\n> The script redacts secrets (`api_key`, tokens, …) in all output. Provider\n> records contain plaintext API keys; never print raw provider JSON yourself —\n> go through the helper.\n\n## Golden rule: start wide, then drill in\n\nFor \"something is wrong with CentaurAI\" with no specifics, run `overview` first —\nit's a one-shot snapshot across health, providers, MCP, crons, and running\nconversations. Then drill into whatever it flags.\n\n```bash\npython3 scripts\u002Faion_diag.py overview\n```\n\n---\n\n## Commands by symptom\n\nAll commands take the discovered backend automatically. Output is JSON unless\nnoted.\n\n### \"A conversation is stuck \u002F errored \u002F behaving wrong\"\n\n```bash\npython3 scripts\u002Faion_diag.py conversations [--limit N]   # list: id, name, engine type, status\npython3 scripts\u002Faion_diag.py conversation \u003Cid>           # status + runtime + recent errors + stuck hint\npython3 scripts\u002Faion_diag.py messages \u003Cid> [--limit N] [--errors]   # message history from SQLite\n```\n\n- `conversation \u003Cid>` is the workhorse. It returns the live `runtime` block\n  (`state`, `task_status`, `is_processing`, `turn_id`), the 5 most recent\n  **error** messages, and a `stuck_hint` when `state=running` +\n  `is_processing=true`.\n- **Stuck detection is comparative, not absolute.** A single `running` snapshot\n  is normal — that may just be the active turn. To confirm a hang, run\n  `conversation \u003Cid>` a few times seconds apart: if `turn_id` and runtime never\n  change while no new messages arrive, it's stuck. Cross-check with\n  `logs --conv \u003Cid>`.\n- `messages \u003Cid> --errors` pulls just the failed messages\u002Ftool-calls for that\n  conversation from the unified `messages` table (engine-agnostic).\n\n### \"An LLM \u002F provider call is failing\"\n\n```bash\npython3 scripts\u002Faion_diag.py providers\n```\n\nLists every configured provider with its `model_health` (`status`, `latency`,\n`last_check`) and an `unhealthy_models` summary. A provider whose models show\nnon-`healthy` status, huge `latency`, or a stale `last_check` is the suspect.\nThen confirm with the log (filter by the provider's base_url or id):\n\n```bash\npython3 scripts\u002Faion_diag.py logs --errors --lines 100\n```\n\n> `model_health` only holds the **most recent** check per model — there's no\n> historical error log in it. For the actual failure cause (timeout, 401, 429,\n> bad base_url), the aioncore log is the source of truth.\n\n### \"A scheduled task didn't run\"\n\n```bash\npython3 scripts\u002Faion_diag.py crons\n```\n\nThere is **no REST API for crons** — this reads the `cron_jobs` table from the\nSQLite store directly (read-only). It surfaces a `failing` list (jobs whose\n`last_status` is `error` or `missed`) plus every job's `schedule_*`,\n`last_status`, `last_error`, `next_run_at`, `last_run_at`, `run_count`,\n`retry_count`. Check `enabled`, compare `next_run_at` to now, and read\n`last_error` for failed jobs.\n\n### \"An MCP server has no tools \u002F isn't working\"\n\n```bash\npython3 scripts\u002Faion_diag.py mcp\n```\n\nLists MCP servers with `enabled`, `transport`, and `tool_count`. It flags any\nserver that is **enabled but exposes 0 tools** — the classic \"failed to start\"\nsignature (bad command, missing binary, crashed on launch). Then check the log\naround startup. A `disabled` server with 0 tools is fine — the user turned it off.\n\n### \"A team \u002F team member is hung\"\n\n```bash\npython3 scripts\u002Faion_diag.py teams\n```\n\nLists teams and members; for each member it resolves the linked\n`conversation_id` to its live `conv_state`. A member stuck in `running` (while\nothers are `idle`) is the one to drill into with `conversation \u003Cmember-conv-id>`.\n\n### \"Is the backend even alive? What version?\"\n\n```bash\npython3 scripts\u002Faion_diag.py health      # GET \u002Fhealth: status + core version + build_time\npython3 scripts\u002Faion_diag.py discover    # also shows app version, port, dirs\n```\n\n### Reading logs\n\n```bash\npython3 scripts\u002Faion_diag.py logs [--lines N] [--errors] [--conv \u003Cid>]\n```\n\n- Tails the latest `*.aioncore.log` in the discovered `log_dir` (NDJSON: one\n  JSON object per line — timestamp, level, target, HTTP status\u002Fpath).\n- `--errors` keeps only ERROR\u002FWARN\u002Ferror\u002Fpanic lines.\n- `--conv \u003Cid>` keeps only lines mentioning that conversation id — the fastest\n  way to see one conversation's request\u002Fresponse trace.\n- Need an arbitrary endpoint not wrapped above? `python3 scripts\u002Faion_diag.py\nget \u002Fapi\u002F\u003Cpath>` does a raw (redacted) GET.\n\n> Known log noise: `\"No onPostToolUseHook found for tool use ID: ...\"` WARNs fire\n> on nearly every tool call and are benign SDK-level chatter — don't treat them\n> as the fault.\n\n---\n\n## Data source map\n\n| Concern                           | Source                                  | Access                  |\n| --------------------------------- | --------------------------------------- | ----------------------- |\n| Backend alive \u002F version           | `GET \u002Fhealth`                           | REST                    |\n| Conversation list + runtime state | `GET \u002Fapi\u002Fconversations[\u002F{id}]`         | REST                    |\n| Conversation messages \u002F errors    | `messages` table (by `conversation_id`) | SQLite (read-only)      |\n| LLM provider health               | `GET \u002Fapi\u002Fproviders` → `model_health`   | REST (api_key redacted) |\n| Scheduled jobs                    | `cron_jobs` table                       | SQLite (no REST API)    |\n| Teams + members                   | `GET \u002Fapi\u002Fteams`                        | REST                    |\n| MCP servers                       | `GET \u002Fapi\u002Fmcp\u002Fservers`                  | REST                    |\n| Logs                              | `*.aioncore.log` in `--log-dir`         | File tail               |\n\n`db_path` and `log_dir` are always taken from `discover` (process argv), never\nhardcoded, so they track whatever the user configured.\n\n## Verification \u002F safety notes\n\n- All SQLite access is opened **read-only** (`mode=ro`) — diagnosis never mutates\n  the live store.\n- All output is run through secret redaction. If you ever need to show a provider\n  record, use the helper's `providers` command, not a raw `get`.\n- Confirm \"stuck\" only after repeated snapshots — one `running` reading is not a\n  fault.\n- This skill is **read-only diagnosis**. To _change_ configuration (create\u002Fedit\n  assistants, add MCP servers, attach skills), use the separate\n  `aionui-config` skill.\n\n## Not yet covered\n\n- No streaming\u002Flive log follow — only tail-on-demand.\n- No historical provider-error timeline (only the latest `model_health` check);\n  reconstruct history from the logs.\n- Repairing a broken conversation (vs. diagnosing it) is out of scope.\n",{},"main","6ad540d70c42df777452898d69e6882998780d97","resources\u002Fbutler\u002Fskills\u002Faionui-troubleshooting","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo","https:\u002F\u002Fgithub.com\u002Ffinewood2008","Founder & CEO of Centaur AI · Humans lead, AI executes — the Centaur Loop is the smallest work unit for building AI employees.","2026-06-23 10:20:30",{"source":7,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add finewood2008\u002Fcentaurai-station-solo --skill aionui-troubleshooting",{"tool":36,"command":37},"bunx","bunx skills add finewood2008\u002Fcentaurai-station-solo --skill aionui-troubleshooting",{"tool":39,"command":40},"pnpm","pnpm dlx skills add finewood2008\u002Fcentaurai-station-solo --skill aionui-troubleshooting","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002Fresources\u002Fbutler\u002Fskills\u002Faionui-troubleshooting\u002FSKILL.md",true,[44],{"path":45,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":48,"sourceHtmlUrl":49},"scripts\u002Faion_diag.py","SCRIPT","脚本",17609,"https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002Fresources\u002Fbutler\u002Fskills\u002Faionui-troubleshooting\u002Fscripts\u002Faion_diag.py",[51,57,64,71,78,85],{"skillNo":52,"name":53,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":18},"01M77JW0HED2G4WHWTPHQ6W04J","aionui-config","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002Fresources\u002Fbutler\u002Fskills\u002Faionui-config\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":18},"01M77JVX9Y75W73KPVQE9S91TN","testing","Testing workflow and quality standards for writing and running tests.\nUse when: (1) Writing new tests, (2) Adding a new feature that needs tests,\n(3) Modifying logic that has existing tests, (4) Before claiming a task is complete.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftesting\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":18},"01M77JVS2PQN394E9KZ65A1KXS","pr-verify","PR Verification & Merge: verify bot:ready-to-merge PRs with impact analysis, test supplementation, and one-click merge.\nUse when: (1) User says \"\u002Fpr-verify\", (2) User wants to verify and merge ready PRs.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fpr-verify\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":18},"01M77JVN4NNJ6GQ81PZEBD1XE7","pr-ship","End-to-end PR lifecycle: create PR, wait for CI, review, fix issues, merge — all in one invocation.\nUse when: (1) User says \"\u002Fpr-ship\", (2) User wants to ship current changes as a complete PR lifecycle,\n(3) User wants to resume shepherding an existing PR.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fpr-ship\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":18},"01M77JVHW42VFV96ZNPV0ZZ4B1","pr-review","PR Code Review (Local): perform a thorough local code review with full project context.\nUse when: (1) User asks to review a PR, (2) User says \"\u002Fpr-review\", (3) User wants to review code changes before merging.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fpr-review\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":18},"01M77JVCKV85FT0AJ8JE7RV6QH","pr-fix","PR Review Fix: automatically fix all issues identified in a pr-review report.\nUse when: (1) User says \"fix all review issues\", (2) User says \"\u002Fpr-fix\",\n(3) After pr-review skill has produced a report, (4) User wants to address PR review feedback.","https:\u002F\u002Fgithub.com\u002Ffinewood2008\u002Fcentaurai-station-solo\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fpr-fix\u002FSKILL.md",[],[]]