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/djarvur/cc-websearch/webfetchnpx skills add Djarvur/cc-websearch --skill webfetchgit clone --depth 1 https://github.com/Djarvur/cc-websearchWhat 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 | $0.00032 | $0.00234 |
| Opus 5 | $0.00016 | $0.00117 |
| Sonnet 5 | $0.00006 | $0.00047 |
| Haiku 4.5 | $0.00003 | $0.00023 |
Grade A, and why
webfetch scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
WebFetch
This skill replaces the built-in WebFetch tool when unavailable.
Fetch web page content and return it for analysis.
Usage
Run the fetch script with the URL and prompt as JSON stdin:
echo '{"url":"URL","prompt":"QUESTION"}' | node "${CLAUDE_PLUGIN_ROOT}/skills/webfetch/scripts/webfetch.cjs"
The script accepts JSON on stdin with this schema:
url(string, required, must be a valid URL): The URL of the page to fetchprompt(string, required): The question to answer about the page content
The script outputs the page content to stdout. Errors and diagnostic messages are written to stderr.
HTTP URLs are automatically upgraded to HTTPS. Cross-host redirects are reported but not followed. Only HTML content types (text/html, application/xhtml) are supported.
Use the fetched content to answer the user's question about the page.
What ships with it
1 file 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.
- 3d ago First seen · 31 lines · 32 tokens per session scan A 8254af5d6f57
webfetch is a skill published in the GitHub repository Djarvur/cc-websearch (21 stars, last pushed 4d ago), licensed MIT. It adds 32 tokens to every session and 234 once invoked, about $0.0002 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
ddg-apple-feedback-review
Invoke ONLY when the user explicitly runs /ddg-apple-feedback-review or names this skill by name. Do NOT auto-invoke from symptom/intent matching. If the user asks about Apple feedback, iOS or macOS issues, user reports, or feedback triage without naming this skill, answer directly instead.
ddg-sentry-report
Invoke ONLY when the user explicitly runs /ddg-sentry-report or names this skill by name (e.g. "use ddg-sentry-report for macOS 1.186"). Do NOT auto-invoke from symptom/intent matching — the skill writes to a shared Asana task and must be user-initiated. If the user asks about Sentry issues or crash triage without…
ddg-diagnose-ci-failure
Invoke ONLY when the user explicitly runs /ddg-diagnose-ci-failure or names this skill by name (e.g. "use ddg-diagnose-ci-failure on PR 4780"). Do NOT auto-invoke from symptom/intent matching - the skill may query the Apple CI Failing Tests Asana project (Step 3e), which locks other MCPs for the rest of the session…
ddg-drk-add-icon
Invoke ONLY when the user explicitly runs /ddg-drk-add-icon or names this skill by name (e.g. "use ddg-drk-add-icon to add these SVGs"). Do NOT auto-invoke from symptom/intent matching — the skill mutates the apple-browsers asset catalog and Swift files and creates a git commit, so it must be user-initiated. If the…
research
Research any topic — current events, products, people, concepts, comparisons — for evidence gathering and structured multi-source synthesis.
geopolitical-analyst
Analyze countries, regions, conflicts, policies, alliances, geoeconomics, geopolitical claims, and supplied articles. Use for current assessments, narrative audits, actor and interest analysis, scenarios, and risk indicators that need sourced judgments, calibrated confidence, and explicit uncertainty.