Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/StarRocks/starrocks-debug-skillsnpx agentmods add skills/starrocks/starrocks-debug-skills/nodeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/starrocks/starrocks-debug-skills/node)<a href="https://agentmods.dev/skills/starrocks/starrocks-debug-skills/node"><img src="https://agentmods.dev/badge/skills/starrocks/starrocks-debug-skills/node/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/starrocks/starrocks-debug-skills/node"><img src="https://agentmods.dev/badge/skills/starrocks/starrocks-debug-skills/node.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00080 | $0.08151 |
| Opus 5.5 | $0.00032 | $0.03260 |
| Sonnet 5.5 | $0.00016 | $0.01630 |
| Haiku 4.5 | $0.00008 | $0.00815 |
Grade A, and why
node scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
description: "Use when a BE or FE node is abnormal: BE out-of-memory or crash, FE deadlock causing all queries to hang, FE Full GC pauses, or FE heap memory growth and leak. Covers mem_tracker curl analysis, large-memory How it starts
The opening of the file, as written. The whole thing — 804 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Node Troubleshooting (BE Crash / OOM, FE Deadlock / GC)
Investigation guide for BE-level crashes and out-of-memory events, FE deadlocks, Full GC pauses, and FE heap memory analysis.
Five root causes account for the vast majority of cases:
- Cause A — Query-caused BE OOM (memory limit exceeded by large query)
- Cause B — BE crash (segfault / internal fatal error)
- Cause C — FE deadlock caused by inconsistent lock ordering
- Cause D — FE Full GC caused by heap too small or memory leak
- Cause E — FE OOM from specific module memory leak
Metric Taxonomy — Read This First
Before using any metrics, understand the two-layer structure:
BE-Side Metrics
| Metric / Source | Meaning |
|---|---|
curl http://<BE>:<port>/mem_tracker |
Per-module memory breakdown (current allocations by subsystem) |
curl http://<BE>:<port>/memz |
tcmalloc allocator internal status — heap vs OS footprint |
starrocks_be_.*_mem_bytes |
Per-module memory gauge in Prometheus/Grafana (e.g., starrocks_be_query_mem_bytes) |
be.WARNING — large memory alloc |
Single large allocation events, includes query_id and stack trace |
dmesg — OOM Killer |
Linux kernel OOM Killer records listing killed process and memory pressure |
FE-Side Metrics
| Metric / Source | Meaning |
|---|---|
fe.gc.log — Full GC entries |
JVM Full GC frequency, stop-the-world pause duration, Old Gen occupancy |
fe.WARNING — FE ShutDown! Running Query |
Queries running at moment of FE OOM shutdown, each with QueryFEAllocatedMemory |
fe.log — LockManager / DeadlockChecker |
Structured JSON deadlock report: lock chain, waiter threads, hold durations |
| Memory Allocated Profile (flame graph) | HTML flame graph in fe/log/proc_profile/ — per-callsite allocation breakdown |
SHOW PROC '/current_queries' |
Live view of running queries with memUsageBytes (FE) and memUsageBytes (BE) |
How to Retrieve Metrics
BE memory breakdown
# Per-module memory (most useful for OOM investigation)
curl -s http://<BE_IP>:<BE_HTTP_PORT>/mem_tracker
# All BE memory gauge metrics
curl -s http://<BE_IP>:<BE_HTTP_PORT>/metrics | grep "^starrocks_be_.*_mem_bytes"
# tcmalloc allocator status
curl -s http://<BE_IP>:<BE_HTTP_PORT>/memz
# Memory-heavy queries from mem_tracker
curl -s http://<BE_IP>:<BE_HTTP_PORT>/mem_tracker | grep "query"
What ships with it
16 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/02-stack-trace-and-memory.md 1.8 KB
- references/03-fe-be-metrics.md 2.1 KB
- references/case-003-fe-deadlock.md 2.6 KB
- references/case-007-memory-tracking-leak.md 2.1 KB
- references/case-008-be-oom.md 2.1 KB
- references/case-021-fe-oom-metadata-operations.md 2.9 KB
- references/case-022-fe-oom-complex-sql-plan.md 3.1 KB
- references/case-023-fe-oom-mv-frequent-refresh.md 3.2 KB
- references/case-024-fe-oom-iceberg-query.md 3.8 KB
- references/case-025-fe-oom-insert-memory-leak.md 2.8 KB
- references/case-026-fe-crash-async-profiler.md 3.7 KB
- references/case-027-fe-oom-heap-config-insufficient.md 4.1 KB
- references/case-029-fe-heap-memory-slow-growth.md 3.2 KB
- references/case-030-fe-machine-memory-growth-malloc-arena.md 3.1 KB
- references/case-031-fe-machine-memory-spike-traffic.md 3.3 KB
- scripts/analyze_logs.py 3.9 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 804 lines · 80 tokens per session scan A 2435083e37f1
node is a skill published in the GitHub repository StarRocks/starrocks-debug-skills (77 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 80 tokens to every session and 8,151 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
doris-debug-node
Use for Doris FE/BE OOM, crash, or false Alive. BE: MemTrackerLimiter lines in be.WARNING; FE: JVM heap / jstack. Distinguish heartbeat Alive from query/brpc serving health.
triage-issues
Triage GitHub issues in the googleapis/mcp-toolbox repo: propose the correct labels (type / priority / product / status), check for duplicates, verify a bug has enough info to act on, and draft a triage comment. Use whenever a maintainer asks you to triage, label, categorize, prioritize, or "look at" an issue (or a…
postgresql-indexing
PostgreSQL indexing best practices for Prowler: index design, partial indexes, partitioned table indexing, EXPLAIN ANALYZE validation, concurrent operations, monitoring, and maintenance. Trigger: When creating or modifying PostgreSQL indexes, analyzing query performance with EXPLAIN, debugging slow queries, reviewing…
graphjin-eval
Create, extend, run, baseline, and diagnose GraphJin agent evaluations through the graphjin eval CLI.
axiom-audit-grdb-performance
Use when the user mentions GRDB performance review, slow GRDB queries, app-group database setup audit, a ValueObservation that stopped updating, or pre-release GRDB scan.
django-perf-review
Django performance code review. Use when asked to "review Django performance", "find N+1 queries", "optimize Django", "check queryset performance", "database performance", "Django ORM issues", or audit Django code for performance problems.