build-list

Name: build-list
Author: chacosoldier

29stars6forks

How to install build-list?

chacosoldier/compabob/build-list

$npx -y skills add chacosoldier/compabob --skill build-list

Installs into the current project.

Summary

Build a ranked, CRM-deduped outbound prospect list from an ICP description. Use for "/build-list", "find me leads matching X", "build a prospect list for Y", or when the user wants a scored outbound list. Orchestrates discovery + the lead-pipeline module + enrichment.

Files · 1

View on GitHub

SKILL.md

1	# Build List
2
3	Turn an ICP description into a ranked, deduped, CRM-aware outbound list using the
4	`lead-pipeline` module. You run the judgement/IO steps (discover, enrich); the
5	script runs the deterministic steps (clean, dedup, score). The cardinal rule:
6
7	> Dedup against the CRM before enriching. Never spend enrichment effort or
8	> credits on accounts the user already knows. Known accounts become warm intros.
9
10	## Preconditions
11
12	- A CRM to dedup against: `data/crm-merge/contacts.json`. If it is missing, tell
13	the user to run `/merge-contacts` first (dedup is the whole point); offer to
14	proceed without it only if they insist, noting every lead will be `proceed`.
15	- An ICP rubric at `modules/lead-pipeline/icp.json`. If only `icp.example.json`
16	exists, copy it and adjust it to the user's request before scoring.
17
18	## Steps
19
20	1. Read the ICP from the request. Pull out: industry/segment, geography,
21	company-size band, and target buyer titles. Confirm anything ambiguous in one
22	short question rather than guessing.
23
24	2. Discover candidates → raw CSV. Use the web-search MCP (e.g. exa
25	`web_search_exa` / `linkedin_search_exa`) to find matching accounts. Write a
26	CSV to `reports/lead-pipeline/00-raw.csv` with whatever you found across these
27	columns: `company, domain, contact_name, contact_email, title, country`
28	(partial rows are fine; later stages fill gaps). If no search MCP is
29	available, ask the user for a raw CSV and skip to step 3.
30
31	3. Clean (script):
32	```bash
33	python3 modules/lead-pipeline/pipeline.py clean \
34	--in reports/lead-pipeline/00-raw.csv \
35	--out reports/lead-pipeline/01-cleaned.csv --country <codes>
36	```
37
38	4. Dedup against the CRM (script):
39	```bash
40	python3 modules/lead-pipeline/pipeline.py dedup \
41	--in reports/lead-pipeline/01-cleaned.csv \
42	--crm data/crm-merge/contacts.json \
43	--out reports/lead-pipeline/02-deduped.csv
44	```
45	Report the split: how many `proceed` (cold), `warm` (known account, route via
46	`crm_known_contact`), `skip` (already known, drop from outbound).
47
48	5. Enrich the survivors only. For rows tagged `proceed` or `warm` that lack a
49	domain / email / decision-maker, enrich with whatever you have, in this order:
50	- web search MCP for the company domain + LinkedIn (high hit-rate, cheap),
51	- an enrichment provider MCP (e.g. Clay) for verified emails — check remaining
52	credits first, and only for rows still missing an email,
53	- public business registries for named decision-makers at small firms when
54	enrichment misses.
55	Never enrich `skip` rows. Write the enriched result back to
56	`reports/lead-pipeline/02-deduped.csv` (or a `02b-enriched.csv`).
57
58	6. Score (script):
59	```bash
60	python3 modules/lead-pipeline/pipeline.py score \
61	--in reports/lead-pipeline/02-deduped.csv \
62	--icp modules/lead-pipeline/icp.json \
63	--out reports/lead-pipeline/03-scored.csv --top 50
64	```
65
66	7. Hand back the ranked top-N (path `reports/lead-pipeline/04-top-*.csv`),
67	the disposition split, and the tier distribution. Call out the `warm`
68	accounts explicitly: those are the highest-value, lowest-risk plays.
69
70	## Guardrails
71
72	- Do not send anything. This builds a list; outreach is a separate, reviewed step
73	(see the `linkedin-outreach` module).
74	- Respect data-protection law and platform terms. Flag if the user's target list
75	looks like scraped personal data being repurposed for cold mail.
76	- If a stage's numbers look wrong (e.g. zero `skip`/`warm` with a populated CRM),
77	inspect the stage CSV before continuing — that is why each stage is a file.
78	- Be honest about enrichment misses; report the hit-rate rather than fabricating
79	emails.

People who liked this also viewed

6 picks

Type

build-list

29stars6forks

View on GitHub

How to install build-list?

chacosoldier/compabob/build-list

$npx -y skills add chacosoldier/compabob --skill build-list

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Build List
2
3	Turn an ICP description into a ranked, deduped, CRM-aware outbound list using the
4	`lead-pipeline` module. You run the judgement/IO steps (discover, enrich); the
5	script runs the deterministic steps (clean, dedup, score). The cardinal rule:
6
7	> Dedup against the CRM before enriching. Never spend enrichment effort or
8	> credits on accounts the user already knows. Known accounts become warm intros.
9
10	## Preconditions
11
12	- A CRM to dedup against: `data/crm-merge/contacts.json`. If it is missing, tell
13	the user to run `/merge-contacts` first (dedup is the whole point); offer to
14	proceed without it only if they insist, noting every lead will be `proceed`.
15	- An ICP rubric at `modules/lead-pipeline/icp.json`. If only `icp.example.json`
16	exists, copy it and adjust it to the user's request before scoring.
17
18	## Steps
19
20	1. Read the ICP from the request. Pull out: industry/segment, geography,
21	company-size band, and target buyer titles. Confirm anything ambiguous in one
22	short question rather than guessing.
23
24	2. Discover candidates → raw CSV. Use the web-search MCP (e.g. exa
25	`web_search_exa` / `linkedin_search_exa`) to find matching accounts. Write a
26	CSV to `reports/lead-pipeline/00-raw.csv` with whatever you found across these
27	columns: `company, domain, contact_name, contact_email, title, country`
28	(partial rows are fine; later stages fill gaps). If no search MCP is
29	available, ask the user for a raw CSV and skip to step 3.
30
31	3. Clean (script):
32	```bash
33	python3 modules/lead-pipeline/pipeline.py clean \
34	--in reports/lead-pipeline/00-raw.csv \
35	--out reports/lead-pipeline/01-cleaned.csv --country <codes>
36	```
37
38	4. Dedup against the CRM (script):
39	```bash
40	python3 modules/lead-pipeline/pipeline.py dedup \
41	--in reports/lead-pipeline/01-cleaned.csv \
42	--crm data/crm-merge/contacts.json \
43	--out reports/lead-pipeline/02-deduped.csv
44	```
45	Report the split: how many `proceed` (cold), `warm` (known account, route via
46	`crm_known_contact`), `skip` (already known, drop from outbound).
47
48	5. Enrich the survivors only. For rows tagged `proceed` or `warm` that lack a
49	domain / email / decision-maker, enrich with whatever you have, in this order:
50	- web search MCP for the company domain + LinkedIn (high hit-rate, cheap),
51	- an enrichment provider MCP (e.g. Clay) for verified emails — check remaining
52	credits first, and only for rows still missing an email,
53	- public business registries for named decision-makers at small firms when
54	enrichment misses.
55	Never enrich `skip` rows. Write the enriched result back to
56	`reports/lead-pipeline/02-deduped.csv` (or a `02b-enriched.csv`).
57
58	6. Score (script):
59	```bash
60	python3 modules/lead-pipeline/pipeline.py score \
61	--in reports/lead-pipeline/02-deduped.csv \
62	--icp modules/lead-pipeline/icp.json \
63	--out reports/lead-pipeline/03-scored.csv --top 50
64	```
65
66	7. Hand back the ranked top-N (path `reports/lead-pipeline/04-top-*.csv`),
67	the disposition split, and the tier distribution. Call out the `warm`
68	accounts explicitly: those are the highest-value, lowest-risk plays.
69
70	## Guardrails
71
72	- Do not send anything. This builds a list; outreach is a separate, reviewed step
73	(see the `linkedin-outreach` module).
74	- Respect data-protection law and platform terms. Flag if the user's target list
75	looks like scraped personal data being repurposed for cold mail.
76	- If a stage's numbers look wrong (e.g. zero `skip`/`warm` with a populated CRM),
77	inspect the stage CSV before continuing — that is why each stage is a file.
78	- Be honest about enrichment misses; report the hit-rate rather than fabricating
79	emails.

People who liked this also viewed

6 picks

Type