Need help implementing this on your team?
We'll install the skills, customize them to your workflow, and run the output through your team's combined LinkedIn network. Skills do the research, Linkd does the motion.
Live
Prioritization
ICP
A list of 300 accounts, ranked top to bottom.
Feed in an account list + an ICP. The skill returns A/B/C tiers with the reasoning behind each row. Differentiation is forced, if 80% of accounts score B, the tiering didn't do its job. Accounts without enough data get flagged "unscored", not guessed.
Differentiation, not confidence-laundered guesses.
Tiering is only useful if the user can act on it. The skill avoids three failures: fake precision (a "97/100 fit score" on partial firmographics), identical scores across the list (if 80% land in B-tier the tiering didn't work), and scoring on dimensions the ICP doesn't actually care about.
The ICP gets translated into weighted scoreable dimensions, industries, headcount range, geographies, technographics, triggers, with weights from either the user or sensible ICP defaults. Each account gets scored across them, with the reasoning per row.
When account data is too thin to score reliably, the row gets flagged "unscored" with a note on what's missing. The user can fill the gap or accept the flag. Either way, no confidence-laundered guesses.
Any of these will work.
Tier this list of 300 accounts against my ICP.
Rank these companies, which should I work first?
A/B/C this list. The ICP is attached.
Score these accounts by fit. CSV pasted below.
Five steps.
01
Normalize the ICP
Translate the ICP into weighted scoreable dimensions. Industries get weight, headcount gets weight, technographics get weight. User can override weights.
02
Enrich what's missing
For accounts with partial firmographics, attempt enrichment via web_search. Flag what couldn't be found rather than guessing.
03
Score each row
Compute a fit score per account. Force differentiation, if too many rows land in B-tier, the skill re-thresholds rather than producing an unhelpful flat list.
04
Assign tiers
A (work first), B (work after A), C (deprioritize), Unscored (insufficient data, surface to user).
05
Emit CSV with reasoning
Each row: tier, score, top contributing dimensions, missing dimensions. The user can audit every assignment.
Built to be auditable, not impressive.
!
No fake precision
"97/100 fit" on partial data is confidence-laundering. The skill produces score bands (A/B/C) with reasoning, not false-precision numerics.
#
Forces differentiation
If 80% of accounts land in B-tier, the tiering failed. The skill re-thresholds to ensure A and C have meaningful population.
@
Unscored is a real tier
Accounts without enough data don't get guessed into a tier. They're flagged "unscored" with a note on what's missing.
?
Every row is auditable
Top contributing dimensions shown per row. The user can spot-check assignments and trust (or correct) the result.
A CSV the user can work top-down.
Each row carries its tier, its fit score, the top 2–3 contributing dimensions, and any missing data. Plus a summary block showing tier distribution and which ICP dimensions had the most differentiating power.
Sample row
Globex Health
domain: globex.io
tier: A
score: 82
strengths: industry match,
headcount match,
recent funding
gaps: technographic
data missing
Tier dist: A: 24, B: 91,
C: 67, ??: 18
Tier contract
- Output
- CSV (tiered) + summary
- Tiers
- A / B / C / Unscored
- Scoring
- ICP-weighted dimensions
- Required tools
- web_search (enrichment)
- Differentiation
- Forced (re-thresholds)
- Pairs with
- icp-extractor
Pick your Claude. Drop in the skill.
The download is a zip containing two things: the .skill file (what you upload to Claude) and a SKILL.md sidecar (the human-readable instructions). Unzip first, then follow the steps for your Claude.
⌘
Most people
Claude.ai or Desktop
- Download and unzip the bundle.
- Open or create a Project in Claude.
- Go to Project knowledge, click Add content.
- Upload the
.skillfile from the unzipped bundle. - Start a new chat in that project. Ask it to run the skill.
Download bundle ↓
No restart needed
›_
Developers
Claude Code
- Download and unzip the bundle.
- Unzip the inner
.skillfile too (it's also a zip). - Move the resulting folder into
~/.claude/skills/ - Restart Claude Code.
- Type a prompt that fits the trigger. The skill loads automatically.
Download bundle ↓
macOS / Linux / WSL
{ }
Builders
API or Agent SDK
- Download and unzip the bundle.
- Open the SKILL.md sidecar to read the full instructions.
- Load via the Skills API, or paste the instructions into your agent's system prompt.
- Make sure your agent has
web_searchand file output enabled. - Ship it.
View SKILL.md ↓
v1.0
Read the full instructions.
This is exactly what Claude reads when the skill triggers. No hidden prompts, no separate config. What you see here is what you get.
account-list-tierer-SKILL.md
⌄
account-list-tierer-SKILL.md
Loading SKILL.md...