--- name: clean-tracking-urls description: Strip tracking and analytics parameters from URLs without breaking the link. Removes utm_*, fbclid, gclid, msclkid, mc_cid, mc_eid, mkt_tok, _hsenc, igshid, ttclid, yclid, twclid, ref, and ~40 other known trackers, while preserving content-essential params like YouTube's v= and t=, search queries, Amazon product IDs, GitHub tabs, and Reddit comment context. Use when the user pastes a long marketing or email URL, asks to clean/tidy/shorten a link before sharing, wants to know what a tracking-lade allowed-tools: Read Edit Write ---
No review checks recorded
Writes files
Runtime not stated
No reviews yet
clean-tracking-urls
Strip tracking and analytics parameters from URLs without breaking the link. Removes utm_*, fbclid, gclid, msclkid, mc_cid, mc_eid, mkt_tok, _hsenc, igshid, ttclid, yclid, twclid, ref, and ~40 other known trackers, while preserving content-essential params like YouTube's v= and t=, search queries, Amazon product IDs,…
- Instant access
- Portable SKILL.md
- Free updates
One file · Writes files
The body stays locked until purchase — here is precisely what unlocks.
Writes files
Declares Read, Edit, Write.
- Writes files Edit, Write
- Reads files Read
What you install
A skill directory you can read the head of before you trust it.
- The SKILL.md file itself1 file · 7.1 KB
- Writes filesRead, Edit, Write
- Free updatesEvery later revision lands in your library automatically.
- SKILL.md 7.1 KB · 2K tk
What it costs to keep loaded
A skill loads in stages — the index entry is always resident, the body only when the skill fires.
Counted with a general-purpose tokenizer, so a particular runtime will differ by a few percent. The skill makes no model call of its own, so there are no output tokens to pay for beyond the reply you were already getting.
Specifics and common questions
- Version
- v1
- Licence
- Not stated
- Allowed tools
- Read, Edit, Write
- Ships as
- 1 file · 7.1 KB
- Runs in
- Any SKILL.md runtime
- Spec
- agentskills.io v1
- Context cost
- 2,000 tokens133 always resident · the rest only when the skill fires.
- Published
- Apr 28, 2026
- Updated
- Aug 1, 2026
- Reviews
- No reviews yet
Experimental bio with extra words, and extra cofe.
-
What does clean-tracking-urls do?
Strip tracking and analytics parameters from URLs without breaking the link. Removes utm_*, fbclid, gclid, msclkid, mc_cid, mc_eid, mkt_tok, _hsenc, igshid, ttclid, yclid, twclid, ref, and ~40 other known trackers, while preserving content-essential params like YouTube's v= and t=, search queries, Amazon product IDs, GitHub tabs, and Reddit comment context. Use when the user pastes a long marketi…
-
Who is clean-tracking-urls best for?
clean-tracking-urls is built for operations & process use cases and is particularly useful for tracking, clean, and urls.
-
How much does clean-tracking-urls cost?
clean-tracking-urls is free to claim on Sigrix.
-
How do I use clean-tracking-urls?
clean-tracking-urls is delivered as a portable SKILL.md module you own on Sigrix. Drop it into Claude Code, claude.ai, or the Claude API and tailor the instructions to your own workflow.
Report this skill
Tell us what's wrong. A moderator reviews every report.
Thanks -- we got it. A moderator will take a look.
More skills
clean-pasted-text
Fix text that was pasted from a PDF, Word doc, webpage, email, or chat and arrived broken. Repairs smart quotes and curly apostrophes, removes non-breaking spaces and zero-width characters, fixes mid-word line breaks and hyphenation from PDF column wraps, normalizes bullets and dashes, collapses double spaces after periods, strips invisible Unicode trash, and rebuilds paragraph structure when newlines were lost or doubled. Use when the user says the text "looks weird," "won't format right," "has
sigrix-listing-workshop
Turn a rough idea, existing prompt, persona, or skill into a clear Sigrix listing. Use when planning, writing, refining, or validating marketplace listing content; preserve the author's intent, identify missing decisions, and avoid unsupported claims.
remove-ai-tells
Strip the structural and lexical patterns that mark text as AI-written, while preserving the author's meaning and any stylistic choices that are genuinely theirs. Targets the real tells — hedge stacking ("it's worth noting that," "it's important to remember"), the "not just X, it's Y" construction, tricolon abuse (three parallel phrases for everything), restating-the-intro conclusions, vague intensifiers ("truly," "deeply," "profoundly"), and overused vocabulary ("delve," "tapestry," "navigate t
ai-ops-setup
AI Ops setup for teams: audit workflows, implement practical AI-agent operations, and deliver a working playbook. Start at https://aiops.neocities.org/ ; payment: https://buy.stripe.com/fZubJ0827bzW8DV02hgUM00