Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/automateyournetwork/netclaw/network-data-analysisnpx skills add automateyournetwork/netclaw --skill network-data-analysisgit clone --depth 1 https://github.com/automateyournetwork/netclawWrote 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/automateyournetwork/netclaw/network-data-analysis)<a href="https://agentmods.dev/skills/automateyournetwork/netclaw/network-data-analysis"><img src="https://agentmods.dev/badge/skills/automateyournetwork/netclaw/network-data-analysis.svg" alt="Measured on agentmods" 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 | $0.00064 | $0.01430 |
| Opus 5 | $0.00032 | $0.00715 |
| Sonnet 5 | $0.00013 | $0.00286 |
| Haiku 4.5 | $0.00006 | $0.00143 |
Grade A, and why
network-data-analysis scanned grade A with 0 findings 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 yesterday.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Network Data Analysis (read-only SQL)
MCP Server
- Server:
analysis-mcp(NetClaw-authored, spec 092) - Tools:
analysis_status,analysis_datasets,analysis_query - Engine: DuckDB, in-memory, sandboxed and locked
What this reads, and what it can never read
Loads files from an allowlist of roots — ~/.openclaw/nsm/runs (Zeek/Suricata output
from nsm-mcp), the workspace output directory, and ~/.openclaw/analysis for your own
exports.
NetClaw's own stores are permanently unreachable: ~/.openclaw/memory/,
~/.openclaw/rag/, ~/.openclaw/n2n/ and ~/.openclaw/gait/, plus .ssh, .aws, .kube
and .env. A generic SQL surface over those would be a backdoor, not an analysis tool.
That is not enforced by pattern matching. Datasets are materialised, then DuckDB's own
enable_external_access=false and lock_configuration=true close every filesystem and
network path irreversibly for the life of the process. Verified: read_csv('/etc/passwd'),
glob('/home/**'), ATTACH of the memory or RAG stores, COPY … TO, INSTALL/LOAD, and
re-enabling access all raise.
If you need a file analysed, put it under ~/.openclaw/analysis — do not try to reach it in
SQL, because you cannot.
Workflow: analyse a capture
- Produce the data first.
nsm_analyze(skillnsm-session-pivot) writes Zeek logs and Suricataeve.json. Without a run, this surface has nothing to read and says so. analysis_status— is the sandbox locked, how many datasets loaded, what are the caps?analysis_datasets— table names, row counts, columns. Read the columns before writing SQL; guessing column names wastes a round trip.analysis_query— one read statement per call.
Zeek tables carry their real column names (id_orig_h, id_resp_p, uid), lifted from
the log's #fields header. Without that they would be column0…columnN and unusable.
The pivot that makes this worth using
Every Zeek log shares uid with conn.log, so a join is the session pivot — and unlike
walking logs one at a time, it aggregates:
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.
- yesterday First seen · 113 lines · 64 tokens per session scan A d537033dc71e
network-data-analysis is a skill published in the GitHub repository automateyournetwork/netclaw (650 stars, last pushed yesterday), licensed Apache-2.0. It adds 64 tokens to every session and 1,430 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
data-analysis
Answer data questions from quick metric lookups to full investigations and stakeholder-ready reports.
Network Security & Traffic Analysis
Network traffic analysis, PCAP parsing, IDS/IPS rule creation, firewall configuration auditing, and network anomaly detection.
hunting-for-cobalt-strike-beacons
Use when detect Cobalt Strike beacon network activity using default TLS certificate signatures (serial 8BB00EE), JA3/JA3S/JARM fingerprints, HTTP C2 profile pattern matching, beacon jitter analysis, and named pipe detection via Zeek, Suricata, and Python PCAP analysis. Use when detecting cobalt strike beacon network…
token-scam-analysis
Deep on-chain scam / rug / soft-rug analysis for EVM tokens (especially Clanker, Doppler, Bankr-style single-admin ERC-20s). Use when the user asks to "analyze this token for scam", "is this a rug", "should I trust this migration", "do a deep dive on holders and deployer", or provides one or more token addresses and…
sql-analysis
Guided workflow for SQL data analysis using dbtools.
duckdb
Use when analytical SQL must run in-process with no server: Parquet/CSV/JSON/Arrow queried in place, OLAP embedded in an app or notebook, a slow pandas groupby on multi-GB data, or S3/lakehouse data read without downloading. NOT a multi-user analytics server (that is clickhouse-analytics), NOT an app's transactional…