Plan, query, sweep. Three Antares CLI commands: profile a repo, investigate one weakness, then check every recommended weakness in parallel. Output as JSON, Markdown or SARIF. That's the lab from Neelesh Pateriya and Barry Yuan.
blogs.cisco.com/cisco-devnet…
1
257
Got an hour? Skip the install.
Pick an interactive NSO Learning Lab, get hands-on with NSO 6.7, and find out what your CLI muscle memory thinks of it. No NSO install needed.
developer.cisco.com/learning…)
3
328
Run your LLM security scan twice. Then diff it, don't re-read it.
In Part 2 of the Security Loop Series, 122 finding records collapsed to 67 stable fingerprints, so the delta is new, recurring or gone. Julio Gomez's tip.
blogs.cisco.com/cisco-devnet…
1
378
Is any SSID in your Meraki estate open right now? A short OPA policy can read each network's SSIDs from the Dashboard API and flag any enabled open one. That's Lori Jakab's example.
blogs.cisco.com/cisco-devnet…
2
383
The first thing AI did on the PubHub rewrite wasn't write code. It read the legacy app and produced an architecture and component design reference. Koushik Dhar tells the story.
blogs.cisco.com/cisco-devnet…
A model invented a CVE citation in a draft security rule. Code review caught it.
So every generated rule ships as a pull request. Julio Gomez shared the miss in Part 3.
blogs.cisco.com/cisco-devnet…
2
452
9 minutes 6 seconds down to 40 seconds, by changing one property name. The tool was fetching full host configuration it never read. Eddie Corban owned the bug and shared the fix.
blogs.cisco.com/cisco-devnet…
8
627
Roughly one in three. In Julio Gomez's 500-repo sweep, Antares put a genuinely vulnerable file in its shortlist 31.7% of the time (hit@any). Cisco's figure for a sweep that size on one H100: about $1. He shows his work.
blogs.cisco.com/cisco-devnet…
The model proposes. A deterministic policy engine decides.
Every tool call gets checked, even reading one file. Matt Shooshtari explains why the language model never gets a vote.
blogs.cisco.com/cisco-devnet…
The best thing the AI asked: what are the routers actually for? That exposed a 91 Gbps forwarding ceiling, and the design became two-tier vPC. Adrian Iliesiu's lesson: ask for a review, not a fact.
blogs.cisco.com/cisco-devnet…
2
462
CSV in. Three ACI tenants out, bridge domains and subnets included. Under three minutes. Jeff Comer does it with Claude on NetGru Ep. 117, and that's only the warm-up.
More from the November issue: ciscodevnet.github.io/newsle… #NetDevOps
Cisco DevNet retweeted
"100% coverage, 0% FOMO." We tested #IoC at #ShiftForward26, where 1,500+ attendees used the AX Pass to deploy agent teams that synthesized 90,000+ insights in real time. Agents didn't just transcribe; they debated and analyzed. See how it came together: cs.co/6011ByTAiT
A late-night emergency change. A temporary exception. Still there.
That's configuration drift, and it's a PCI DSS 4.0.1 Requirement 2 problem. William Nellis shows how Cisco IQ compares live configs to approved standards.
blogs.cisco.com/cisco-devnet…
3
472
One line in your Flink config turns on the flame graph:
rest.flamegraph.enabled: true
It's opt-in. For Soumitra Kumar and Annie Zhou, it pointed right at the culprit.
blogs.cisco.com/cisco-devnet…
474
Somebody wrote the page that got you through your first kubectl apply.
Natali Vlatko co-chairs Kubernetes SIG Docs, the group behind those pages. Thank you, Natali, and congrats again on the CNCF Lorem Ipsum Award. #Kubernetes
📸 Gary Kevorkian & friends
5
512
🌧️ Rain, a storm and a whole lot of open source! Poncho Sandoval’s AI-packed week in Lisbon: a talk for tech leaders, a Web Summit masterclass (111 people, room for 100 😅) and one big lesson: record your demos 🎬
Read more: 🔗 cs.co/9009ByT9b5
#AI #WebSummit #DevNet
3
498
More than 20 skills, 2 review agents, 7 practices. And a human still approves every change.
Neelesh Pateriya shares how the PubHub team turned what worked into a reusable AI-first SDLC playbook, in Part 2.
blogs.cisco.com/cisco-devnet…
433
The trick: test your telnet guardrail like any other code. Write the Rego rule for VTY transport input, add an inline test input, run opa test. Lori Jakab shows the whole loop.
blogs.cisco.com/cisco-devnet…
2
569
"Working" is not "finished."
The AI-built PubHub app ran end to end, yet the visual language was off and workflows needed cleanup. Thanks to Alejandro Marrufo, Casey Tong and Jerry Chang for sharing Part 3.
blogs.cisco.com/cisco-devnet…
4
531
The honest cost of verified LLM findings: about 10x the tokens of a single-prompt baseline on Langflow.
Julio Gomez put the price in the open. Know the trade before you sweep (Security Loop Series, Part 2).
blogs.cisco.com/cisco-devnet…
7
639



















