Stats
Actions
Tags
From neptune-dxp
Read Neptune DXP system log files via the MCP tools `list_system_logs` and `get_system_log`. Use when the user wants to debug a server error, inspect script output, trace requests, or investigate exceptions/vault activity from the platform's daily log files. Trigger phrases include "check the logs", "show server errors", "what did the script log", "any exceptions today", "tail the request log", "look at the system log", "debug this error in the logs".
How this skill is triggered — by the user, by Claude, or both
Slash command
/neptune-dxp:inspect-system-logsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Planet 9 writes its operational logs to daily files — **one file per day per category**. These two read-only tools let you find those files and read entries from them. This is the go-to for debugging runtime errors, seeing what a server script logged, or tracing requests.
Planet 9 writes its operational logs to daily files — one file per day per category. These two read-only tools let you find those files and read entries from them. This is the go-to for debugging runtime errors, seeing what a server script logged, or tracing requests.
type | What it holds |
|---|---|
server | General server/runtime log. |
exceptions | Uncaught exceptions and stack traces. |
scripts | Server-script output (log.<silly, verbose, info, debug, warn, error, crit>, e.g log.debug("User not framed")), console.log does not go to this log |
requests | Incoming HTTP request log. |
vault | Secret/vault access activity. |
| Tool | Purpose |
|---|---|
list_system_logs({ type?, dateFrom?, dateTo? }) | List available log files. Omit type to get files grouped across all categories; pass one to narrow. |
get_system_log({ type, name, ... }) | Read entries from a specific file (paginated, filterable). |
list_system_logs returns files grouped as serverFiles/exceptionFiles/scriptFiles/requestFiles/vaultFiles (or pass type to narrow); names look like 2026-05-11-planet9.log. get_system_log needs type + that name, and filters by: level (silly→crit), message (case-insensitive substring), scriptName/scriptGroup (scripts category only), start/end (default 0–100, end exclusive), dateFrom/dateTo.
name must come from list_system_logs — don't guess; the naming convention varies by deployment. So the flow is always list → get.level and/or message filter when hunting a problem. Files are huge; an unfiltered 0–100 window from a busy requests log tells you little. Filter to level: ["error","crit"] or a distinctive message substring.syslog role (List/Get).run-server-script / manage-server-scripts — when the scripts log points at a failing script, inspect and re-run it.manage-apis — request/proxy failures in the requests log often trace back to an API artifact.search-docs — look up an unfamiliar error message in the public DXP docs.npx claudepluginhub neptune-software/claude-code-plugin --plugin neptune-dxpCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.