Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add MikkoParkkola/nab/plugin install nabWrote 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/mikkoparkkola/nab/url-insight)<a href="https://agentmods.dev/skills/mikkoparkkola/nab/url-insight"><img src="https://agentmods.dev/badge/skills/mikkoparkkola/nab/url-insight/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/mikkoparkkola/nab/url-insight"><img src="https://agentmods.dev/badge/skills/mikkoparkkola/nab/url-insight.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 | $0.00127 | $0.12971 |
| Opus 5 | $0.00063 | $0.06485 |
| Sonnet 5 | $0.00025 | $0.02594 |
| Haiku 4.5 | $0.00013 | $0.01297 |
Grade A, and why
url-insight 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 10d 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.
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 — 874 lines — stays where its author put it; the contents beside it link to each section on GitHub.
url-insight — URL batches → portfolio insights → DoR issues
Fires on: multi-URL paste · "analyze these urls" · "should we adopt X" · "competitive signal" · "worth a spike" · arxiv/HN/ProductHunt/GitHub URLs in bulk. Purpose: compress a firehose of URLs into ranked, evidence-backed actions against the 21-active-repo portfolio. Noise → IGNORE. Signal → specialist agent panel → Linear ticket with devil + steelman reckoning.
Design: Fukasawa (paste URL → pipeline fires) · Rams #10 (8 labels, not 30) · Honest (IGNORE is first-class) · Calibrated (devil + steelman together).
0 · MANDATORY protocol (non-negotiable, like hebb/linear)
- Before fetching any URL: run the dedup pre-check in §0.1. If hebb or Linear finds a prior verdict, skip fetch, skip nab, and report
Already triaged: <MIK-id>, verdict=<ADOPT|IGNORE|INSPIRE|SUPERCHARGE|HUMAN-REVIEW>, ROI=<n>xwith the issue link. - After each URL is classified + scored: write the hebb dedup pin in §0.1 with
project="url-insight"and key schemaurl-<sha256-of-canonical-url>-<iso-date>. - Before filing any issue: run devil's advocate (§6) AND steelman (§6.5). Quote both verbatim in issue body.
- For SUPERCHARGE: dispatch innovator agent FIRST (serial gate). If kill → drop signal. If proceed → parallel fan-out: moonshot-architect + business-panel + research patent-prior-art.
- For FAIL-FAST SPIKE: surface user prompt "file issue / also spawn <1h spike agent?" (unless
--auto-spikeflag). - Emit innovation-tracker queue entry per actionable insight (§10).
- Wayback evidence capture (BLOCKING for non-IGNORE): after a successful fetch, call
fulcrum:wayback_savewithif_not_archived_within="7d"andcapture_screenshot="1"; persistwayback_urlorwayback_save_failed=truein the run JSON (§1.1). Capture failure is evidence debt, not a triage blocker. - Dual output: JSON artifact
~/.claude/data/url-insight/run-TIMESTAMP.json+ markdown report. - Living-document update (BLOCKING for non-IGNORE): append delta line(s) with
[^N]citation(s) to~/.claude/data/portfolio/market-positioning-living.mdper §0.8. No skipping. No "I'll do it later". The doc rots without per-URL writes.
What ships with it
3 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.
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.
- 10d ago First seen · 874 lines · 127 tokens per session scan A e0e1192bd61c
url-insight is a skill published in the GitHub repository MikkoParkkola/nab (12 stars, last pushed yesterday), licensed MIT. It adds 127 tokens to every session and 12,971 once invoked, about $0.0006 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 2026-08-30.
Other skills, from other repositories
feishu
Work with Feishu or Lark bots, docs, sheets, bitables, approval flows, and OpenAPI/MCP setup without hardcoding credentials.
Agent Browser Automation
Fast Rust-based headless browser automation CLI with Node.js fallback for AI agents, featuring navigation, clicking, typing, snapshots, and structured commands optimized for agent workflows.
reflect
Use Recall Reflect to review AI coding workflow history as a conversation-first timeline and discuss observed patterns before changing behavior. Trigger when the user asks to reflect on a project, review AI coding workflow, inspect workflow friction, discuss repeated corrections, review handoffs, or explore possible…
text-processing
Transform, filter, sort, and analyze text using sed, awk, sort, uniq, cut, tr, wc, and other text processing tools. Use when the user asks to find and replace text in files, extract columns from data, count lines or words, sort and deduplicate output, reformat delimited data, perform field-based calculations, or build…
scheduler
Create, cancel, and manage periodic (cron) and one-shot (deferred) background tasks. Use when the user wants to schedule recurring work (daily summaries, cleanups), run something at a specific time, set reminders, or manage existing scheduled tasks. Keywords: schedule, cron, timer, remind, recurring, periodic…
query
Answer questions about the wiki by searching, reading, and synthesizing relevant scraps with [[Title]] citations. Use this when the user wants to query the wiki, ask what they have written about a topic, compare scraps, find related notes, or pull a synthesized overview from existing scraps.