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/taboola/realize-claude-plugin/web-fallbacknpx skills add taboola/realize-claude-plugin --skill web-fallbackgit clone --depth 1 https://github.com/taboola/realize-claude-pluginWhat 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 | $0.00099 | $0.02057 |
| Opus 5 | $0.00049 | $0.01028 |
| Sonnet 5 | $0.00020 | $0.00411 |
| Haiku 4.5 | $0.00010 | $0.00206 |
Grade A, and why
web-fallback 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 2d 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Web Fallback
The knowledge base is finite; the questions aren't. There are 12 topic files, and an advertiser can ask hundreds of reasonable "how do I…" questions that fall just outside them. Before this skill existed, those questions got a thin answer from model priors or a shrug — neither of which is what the user came for, and the first is worse than the second.
This skill closes that gap with one narrow move: look the topic up in Taboola's public advertiser help documentation, and answer with a "this is what I found online" framing. It is the last tier of the sourcing ladder in os/guardrails.md → Sourcing — prioritize the plugin's own sources, and it never outranks anything above it.
When to use
Trigger only when all four hold:
- The question is about advertising on Realize — a setup step, a platform feature, a policy, a how-to, a term the user hit in the UI.
- No
knowledge/topic file answers the question as asked. Checkknowledge/manifest.jsonfirst, every time. - No MCP tool answers it. This isn't an account, campaign, item, catalog, or report question.
- It isn't a refusal case. Out-of-scope requests (forecasting, ROI projections, competitor comparisons, legal advice, creative copywriting, landing-page critique) and UI-only actions keep the answers they have today.
Not for adding polish to an answer the plugin can already give. That's the failure mode this skill is one bad judgment call away from: once web content starts topping up good answers, the plugin's own curated guidance stops being the thing users hear, and the freshness problem it exists to avoid arrives anyway.
The miss test — per question, not per topic
The topic file existing is not coverage. What matters is whether it answers the question that was asked:
| Situation | Do |
|---|---|
| The knowledge base answers the question | Answer from it. No lookup. |
| The knowledge base covers the topic but is silent on the actual ask | Miss — look it up. |
| The knowledge base answers it; a lookup would only add detail | Knowledge base wins. No lookup. |
| Out of scope, or a UI-only action | Existing refusal / UI redirect, unchanged. No lookup. |
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.
- 2d ago First seen · 88 lines · 99 tokens per session scan A 633f63188c5e
web-fallback is a skill published in the GitHub repository taboola/realize-claude-plugin (3 stars, last pushed 2d ago), licensed Apache-2.0. It adds 99 tokens to every session and 2,057 once invoked, about $0.0005 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 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
agent-host-chat-contributions
Build and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.