Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

npm Downloads to Leads · varnan-tech | DecimalAI

Sales & Supportby varnan-tech

npm Downloads to Leads

Takes a list of npm package names (yours or competitors'), fetches 12 weeks of daily download data from the npm API, computes a breakout velocity score per package to identify hockey-stick growth, fetches maintainer profiles from the npm registry and GitHub API, and outputs a ranked lead brief for each breakout package with who built it, how to reach them, and what to say.

Maintained

SkillScore

SkillScore64/100

Test results 64

Safety

3/3 safety checks passed

Code scan
Intent
Content

What it changes

27%64%of test cases passed +36 pts

Without skill6 of 22

With skill14 of 22

9 fixed5 passed either way7 still failing1 broke

Measured on gemini-3.6-flash8/18/2026v119 comparable

Track record

Agent runs: 0; All time

See one task, with and without the skill

One test taskWhen enriching maintainer contacts, the GitHub API user response returns a null `twitter_username` field for an npm maintainer. What explicit string value should be assigned to the twitter contact field?

Without the skillMissed

With the skillPassed

Same model, same task. Answers are capped and redacted.

Use this skill

Your agent runs on

Adds every file the skill has, pinned so you can update it. No account needed.

pip install "decimalai>=0.15"
decimalai skills add varnan-tech-npm-downloads-to-leads
claude --plugin-dir "$(git rev-parse --show-toplevel 2>/dev/null || pwd)/.agents"

The same install from Python, for a notebook or a setup script.

Install

pip install "decimalai>=0.15"

Add the skill

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

print("Added. Start Claude Code with the command below.")

Start Claude Code with it

claude --plugin-dir "$(git rev-parse --show-toplevel 2>/dev/null || pwd)/.agents"

Type these in a Claude Code session. No account or API key. It writes the skill and any files it bundles, and asks first when its safety rating isn't Passed.

Add the DecimalAI plugin, once

/plugin marketplace add decimal-labs/decimalai-claude-plugins
/plugin install decimalai@decimalai

Install this skill

/decimalai:install varnan-tech-npm-downloads-to-leads

The plugin needs python3 and curl on your machine.

No Python or account. Claude Code picks it up in its next session.

Save it into your project

curl -fsSL --create-dirs -o .claude/skills/varnan-tech-npm-downloads-to-leads/SKILL.md https://app.decimal.ai/s/varnan-tech-npm-downloads-to-leads/SKILL.md

Copy SKILL.md

Save it in your project, as

.claude/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Claude Code docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[claude-agent-sdk]"
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

import asyncio

from claude_agent_sdk import ClaudeAgentOptions
from decimalai.claude_agent_sdk import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

async def main():
    question = "Hello!"
    async with support.run(question) as run:  # one user message, one run
        async for message in run.query(options=ClaudeAgentOptions(max_turns=6)):
            print(message)

# In a notebook or async code: await main()
asyncio.run(main())

Prefer its own loader? ClaudeAgentOptions(plugins=[skills_plugin()], skills="all") reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

The same install from Python, for a notebook or a setup script. agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[claude-agent-sdk]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

import asyncio

from claude_agent_sdk import ClaudeAgentOptions
from decimalai.claude_agent_sdk import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

async def main():
    question = "Hello!"
    async with support.run(question) as run:  # one user message, one run
        async for message in run.query(options=ClaudeAgentOptions(max_turns=6)):
            print(message)

# In a notebook or async code: await main()
asyncio.run(main())

Prefer its own loader? ClaudeAgentOptions(plugins=[skills_plugin()], skills="all") reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

import asyncio
from pathlib import Path
from claude_agent_sdk import ClaudeAgentOptions, query

skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
options = ClaudeAgentOptions(system_prompt=f"You are a helpful agent.\n\n{skill}", max_turns=6)

async def main():
    async for message in query(prompt="Hello!", options=options):
        print(message)

# In a notebook or async code: await main()
asyncio.run(main())

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Claude Agent SDK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[openai-agents]"
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

from agents import Agent, Runner
from decimalai.openai_agents import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    name=support.name,
    instructions=support.instructions,  # your prompt, then the skill list
    tools=[*support.tools],             # your tools, then load_skill
)
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

Using SandboxAgent? capabilities=[Shell(), decimalai.openai_agents.sandbox_skills()] points its own loader at the same folder.

Install

pip install "decimalai[openai-agents]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

from agents import Agent, Runner
from decimalai.openai_agents import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    name=support.name,
    instructions=support.instructions,  # your prompt, then the skill list
    tools=[*support.tools],             # your tools, then load_skill
)
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

Using SandboxAgent? capabilities=[Shell(), decimalai.openai_agents.sandbox_skills()] points its own loader at the same folder.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from agents import Agent, Runner

skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
agent = Agent(name="support", instructions=f"You are a helpful agent.\n\n{skill}")
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

OpenAI Agents docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[openai]"
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

import json

from openai import OpenAI
from decimalai.openai import agent_config

client = OpenAI()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your tools, as Chat Completions takes them, and what answers each.
TOOLS = [{"type": "function", "function": {
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters": {"type": "object", "properties": {"order_id": {"type": "string"}},
                   "required": ["order_id"]},
}}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    # run.system is your prompt, then the skill list.
    messages = [{"role": "system", "content": run.system}, run.message]
    while True:
        reply = client.chat.completions.create(
            model="gpt-5.4-mini",
            messages=messages,
            tools=[*TOOLS, *run.tools],  # your tools, then load_skill
        ).choices[0].message
        messages.append(reply)
        if not reply.tool_calls:
            break
        for call in reply.tool_calls:
            # load_skill is answered by the run; your own tools by FUNCTIONS.
            function = FUNCTIONS.get(call.function.name)
            messages.append(run.tool_result(call) or {
                "role": "tool", "tool_call_id": call.id,
                "content": function(**json.loads(call.function.arguments)) if function
                else f"There is no tool named {call.function.name!r}.",
            })
print(reply.content)

# The Responses API takes run.response_tools instead of run.tools.

Install

pip install "decimalai[openai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

import json

from openai import OpenAI
from decimalai.openai import agent_config

client = OpenAI()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your tools, as Chat Completions takes them, and what answers each.
TOOLS = [{"type": "function", "function": {
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters": {"type": "object", "properties": {"order_id": {"type": "string"}},
                   "required": ["order_id"]},
}}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    # run.system is your prompt, then the skill list.
    messages = [{"role": "system", "content": run.system}, run.message]
    while True:
        reply = client.chat.completions.create(
            model="gpt-5.4-mini",
            messages=messages,
            tools=[*TOOLS, *run.tools],  # your tools, then load_skill
        ).choices[0].message
        messages.append(reply)
        if not reply.tool_calls:
            break
        for call in reply.tool_calls:
            # load_skill is answered by the run; your own tools by FUNCTIONS.
            function = FUNCTIONS.get(call.function.name)
            messages.append(run.tool_result(call) or {
                "role": "tool", "tool_call_id": call.id,
                "content": function(**json.loads(call.function.arguments)) if function
                else f"There is no tool named {call.function.name!r}.",
            })
print(reply.content)

# The Responses API takes run.response_tools instead of run.tools.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from openai import OpenAI

client = OpenAI()
skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
reply = client.chat.completions.create(
    model="gpt-5.4-mini",
    messages=[
        {"role": "system", "content": f"You are a helpful agent.\n\n{skill}"},
        {"role": "user", "content": "Hello!"},
    ],
).choices[0].message

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

OpenAI SDK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[anthropic]"
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

import anthropic
from decimalai.anthropic import agent_config

client = anthropic.Anthropic()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

question = "Hello!"
with support.run(question) as run:  # one user message, one run
    answer = client.beta.messages.tool_runner(
        model="claude-haiku-5-5",
        max_tokens=4096,
        system=run.system,     # your prompt, then the skill list
        tools=[*run.tools],    # your tools, then load_skill
        messages=[{"role": "user", "content": question}],
    ).until_done()

Install

pip install "decimalai[anthropic]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

import anthropic
from decimalai.anthropic import agent_config

client = anthropic.Anthropic()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

question = "Hello!"
with support.run(question) as run:  # one user message, one run
    answer = client.beta.messages.tool_runner(
        model="claude-haiku-5-5",
        max_tokens=4096,
        system=run.system,     # your prompt, then the skill list
        tools=[*run.tools],    # your tools, then load_skill
        messages=[{"role": "user", "content": question}],
    ).until_done()

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
import anthropic

client = anthropic.Anthropic()
skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
reply = client.messages.create(
    model="claude-haiku-5-5",
    max_tokens=4096,
    system=f"You are a helpful agent.\n\n{skill}",
    messages=[{"role": "user", "content": "Hello!"}],
)

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Anthropic docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[google-genai]"
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

from google import genai
from google.genai import types
from decimalai.google_genai import agent_config

client = genai.Client()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your functions, as declarations for the model, and what answers each.
TOOLS = [{"function_declarations": [{
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters_json_schema": {"type": "object", "properties": {"order_id": {"type": "string"}},
                               "required": ["order_id"]},
}]}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    config = types.GenerateContentConfig(
        system_instruction=run.system,     # your prompt, then the skill list
        tools=[*TOOLS, *run.tool_params],  # your functions, then load_skill
    )
    contents = [{"role": "user", "parts": [{"text": run.message}]}]
    while True:
        response = client.models.generate_content(
            model="gemini-3.5-flash", contents=contents, config=config,
        )
        if not response.function_calls:
            break
        contents.append(response.candidates[0].content)
        parts = []
        for call in response.function_calls:
            # load_skill is answered by the run; your own functions by FUNCTIONS.
            # A loop of your own rather than automatic function calling, which
            # raises KeyError when the model asks for a function you did not pass.
            function = FUNCTIONS.get(call.name)
            parts.append(run.tool_result(call) or {"function_response": {
                "name": call.name, "id": call.id,
                "response": {"result": function(**(call.args or {})) if function
                             else f"There is no function named {call.name!r}."},
            }})
        contents.append({"role": "user", "parts": parts})
print(response.text)

Install

pip install "decimalai[google-genai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

from google import genai
from google.genai import types
from decimalai.google_genai import agent_config

client = genai.Client()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your functions, as declarations for the model, and what answers each.
TOOLS = [{"function_declarations": [{
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters_json_schema": {"type": "object", "properties": {"order_id": {"type": "string"}},
                               "required": ["order_id"]},
}]}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    config = types.GenerateContentConfig(
        system_instruction=run.system,     # your prompt, then the skill list
        tools=[*TOOLS, *run.tool_params],  # your functions, then load_skill
    )
    contents = [{"role": "user", "parts": [{"text": run.message}]}]
    while True:
        response = client.models.generate_content(
            model="gemini-3.5-flash", contents=contents, config=config,
        )
        if not response.function_calls:
            break
        contents.append(response.candidates[0].content)
        parts = []
        for call in response.function_calls:
            # load_skill is answered by the run; your own functions by FUNCTIONS.
            # A loop of your own rather than automatic function calling, which
            # raises KeyError when the model asks for a function you did not pass.
            function = FUNCTIONS.get(call.name)
            parts.append(run.tool_result(call) or {"function_response": {
                "name": call.name, "id": call.id,
                "response": {"result": function(**(call.args or {})) if function
                             else f"There is no function named {call.name!r}."},
            }})
        contents.append({"role": "user", "parts": parts})
print(response.text)

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from google import genai
from google.genai import types

client = genai.Client()
skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
response = client.models.generate_content(
    model="gemini-3.5-flash",
    contents="Hello!",
    config=types.GenerateContentConfig(system_instruction=f"You are a helpful agent.\n\n{skill}"),
)

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Gemini API docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[langchain]" langchain-openai
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

from langchain.agents import create_agent
from decimalai.langchain import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = create_agent(
    "openai:gpt-5.4-mini",
    tools=[],                         # your tools; the middleware adds load_skill
    middleware=[support.middleware],  # your prompt, the skill list and load_skill
)
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

Install

pip install "decimalai[langchain]" langchain-openai

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

from langchain.agents import create_agent
from decimalai.langchain import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = create_agent(
    "openai:gpt-5.4-mini",
    tools=[],                         # your tools; the middleware adds load_skill
    middleware=[support.middleware],  # your prompt, the skill list and load_skill
)
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from langchain.agents import create_agent

skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
agent = create_agent("openai:gpt-5.4-mini", tools=[], system_prompt=f"You are a helpful agent.\n\n{skill}")
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

LangChain docs

CrewAI's own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" "crewai>=1.15.27"
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

import decimalai.crewai
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                               # your own CrewAI LLM
    skills=decimalai.crewai.skills(),      # the installed folder
    tools=decimalai.crewai.skill_tools(),  # reads a skill's files, runs its scripts
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

The same install from Python, for a notebook or a setup script. CrewAI's own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" "crewai>=1.15.27"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

import decimalai.crewai
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                               # your own CrewAI LLM
    skills=decimalai.crewai.skills(),      # the installed folder
    tools=decimalai.crewai.skill_tools(),  # reads a skill's files, runs its scripts
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

No DecimalAI package. CrewAI's own loader reads the folder, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                         # your own CrewAI LLM
    skills=[Path(".agents/skills")],
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

CrewAI docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[adk]"
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

from decimalai.adk import DecimalAIPlugin, agent_config
from google.adk.agents import LlmAgent
from google.adk.apps import App

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = LlmAgent(
    name="support",          # ADK needs a Python identifier
    model="gemini-3.5-flash",
    instruction=support.instruction,  # your prompt, then the skill list
    tools=[*support.tools],           # your tools, then load_skill
)
# DecimalAIPlugin makes each invocation of runners built from this App one trace.
app = App(name="support", root_agent=agent, plugins=[DecimalAIPlugin()])

Prefer ADK's own loader? SkillToolset(skills=decimalai.adk.skills(), ...) reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

Install

pip install "decimalai[adk]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

from decimalai.adk import DecimalAIPlugin, agent_config
from google.adk.agents import LlmAgent
from google.adk.apps import App

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = LlmAgent(
    name="support",          # ADK needs a Python identifier
    model="gemini-3.5-flash",
    instruction=support.instruction,  # your prompt, then the skill list
    tools=[*support.tools],           # your tools, then load_skill
)
# DecimalAIPlugin makes each invocation of runners built from this App one trace.
app = App(name="support", root_agent=agent, plugins=[DecimalAIPlugin()])

Prefer ADK's own loader? SkillToolset(skills=decimalai.adk.skills(), ...) reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from google.adk.agents import LlmAgent

skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
agent = LlmAgent(
    name="support",
    model="gemini-3.5-flash",
    instruction=lambda context: f"You are a helpful agent.\n\n{skill}",
)

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Google ADK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[pydantic-ai]"
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

from pydantic_ai import Agent
from decimalai.pydantic_ai import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    "openai:gpt-4o-mini",
    name=support.name,
    tools=[],                 # your tools; the capability adds load_skill
    capabilities=[support],   # your prompt, the skill list and load_skill
)
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

On the Pydantic AI harness, decimalai.pydantic_ai_harness.skills(project) uses its own loader instead.

Install

pip install "decimalai[pydantic-ai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

from pydantic_ai import Agent
from decimalai.pydantic_ai import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    "openai:gpt-4o-mini",
    name=support.name,
    tools=[],                 # your tools; the capability adds load_skill
    capabilities=[support],   # your prompt, the skill list and load_skill
)
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

On the Pydantic AI harness, decimalai.pydantic_ai_harness.skills(project) uses its own loader instead.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from pydantic_ai import Agent

skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
agent = Agent("openai:gpt-4o-mini", instructions=f"You are a helpful agent.\n\n{skill}")
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Pydantic AI docs

Deep Agents' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" deepagents
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

import decimalai.deepagents
from deepagents import create_deep_agent

agent = create_deep_agent(model=model, **decimalai.deepagents.skills())
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

The default backend reads files but has no shell. To run a skill's scripts, pass skills(backend=LocalShellBackend(...)).

The same install from Python, for a notebook or a setup script. Deep Agents' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" deepagents

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

import decimalai.deepagents
from deepagents import create_deep_agent

agent = create_deep_agent(model=model, **decimalai.deepagents.skills())
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

The default backend reads files but has no shell. To run a skill's scripts, pass skills(backend=LocalShellBackend(...)).

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from deepagents import create_deep_agent

skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
agent = create_deep_agent(model=model, system_prompt=f"You are a helpful agent.\n\n{skill}")
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Deep Agents docs

Its own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" agent-framework
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

import asyncio

from agent_framework import Agent
from decimalai.agent_framework import skills

agent = Agent(client=client, instructions="You are a helpful agent.", context_providers=[skills()])
asyncio.run(agent.run("Hello!"))

It caches skills by default, so restart the process after decimalai skills update.

The same install from Python, for a notebook or a setup script. Its own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" agent-framework

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

import asyncio

from agent_framework import Agent
from decimalai.agent_framework import skills

agent = Agent(client=client, instructions="You are a helpful agent.", context_providers=[skills()])
asyncio.run(agent.run("Hello!"))

It caches skills by default, so restart the process after decimalai skills update.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

import asyncio
from pathlib import Path
from agent_framework import Agent

skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
agent = Agent(client=client, instructions=f"You are a helpful agent.\n\n{skill}")
asyncio.run(agent.run("Hello!"))

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Microsoft Agent Framework docs

Strands' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" strands-agents
decimalai skills add varnan-tech-npm-downloads-to-leads

Add it to your agent

from decimalai.strands import skills
from strands import Agent
from strands.vended_tools import file_editor, shell

agent = Agent(
    model=model,                # always pass one: the default is Amazon Bedrock
    system_prompt="You are a helpful agent.",
    tools=[file_editor, shell],
    plugins=[skills()],
)
agent("Hello!")

Strands loads an agent's skills once, so build the agent again after decimalai skills update.

The same install from Python, for a notebook or a setup script. Strands' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" strands-agents

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("varnan-tech-npm-downloads-to-leads")
if result.exit_code:
    raise SystemExit(result.problems)

from decimalai.strands import skills
from strands import Agent
from strands.vended_tools import file_editor, shell

agent = Agent(
    model=model,                # always pass one: the default is Amazon Bedrock
    system_prompt="You are a helpful agent.",
    tools=[file_editor, shell],
    plugins=[skills()],
)
agent("Hello!")

Strands loads an agent's skills once, so build the agent again after decimalai skills update.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

Give it to your agent

from pathlib import Path
from strands import Agent

skill = Path(".agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md").read_text()
agent = Agent(model=model, system_prompt=f"You are a helpful agent.\n\n{skill}")
agent("Hello!")

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Strands docs

Adds every file the skill has, pinned so you can update it. Cursor, Copilot, Cline and Warp read this folder by themselves.

pip install "decimalai>=0.15"
decimalai skills add varnan-tech-npm-downloads-to-leads

No Python or account. Cursor, Copilot, Cline and Warp read this folder by themselves.

Save it into your project

curl -fsSL --create-dirs -o .agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md https://app.decimal.ai/s/varnan-tech-npm-downloads-to-leads/SKILL.md

Copy SKILL.md

Save it in your project, as

.agents/skills/varnan-tech-npm-downloads-to-leads/SKILL.md

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Cursor and other editors docs

Fetch the file and give its text to your model as instructions. No account or package needed.

Get the file

curl -fsSL https://app.decimal.ai/s/varnan-tech-npm-downloads-to-leads/SKILL.md

Read the summary as JSON

curl -fsSL https://app.decimal.ai/s/varnan-tech-npm-downloads-to-leads.json

This gets SKILL.md only. The skill's 3 other files are not included.

Other ways to add it

Other setups docs

You don't need a DecimalAI account.

Version: v1, tested 8/18/2026
Source: varnan-tech/opendirectory ↗
License: MIT

Manage it in DecimalAI

Saved prompts and skills, pinned versions, and traces of this skill in your runs. DecimalAI does not run your agent.

Claude Code reads files from your project, so DecimalAI config does not apply there.

On Claude Agent SDK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On OpenAI Agents: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On OpenAI SDK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Anthropic: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Gemini API: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On LangChain: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On CrewAI: set DECIMAL_API_KEY and add decimalai.init(crewai=True) to record runs and skill use. The skill version stays in your repo.

On Google ADK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Pydantic AI: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Deep Agents: set DECIMAL_API_KEY and add decimalai.init(langchain=True) to record runs and skill use. The skill version stays in your repo.

On Microsoft Agent Framework: set DECIMAL_API_KEY and add decimalai.init(otel=True, agent_name="support") to record runs and skill use. The skill version stays in your repo.

On Strands: set DECIMAL_API_KEY and add decimalai.init(otel=True, agent_name="support") to record runs and skill use. The skill version stays in your repo.

Editors read files from your project, so DecimalAI config does not apply there.

Anywhere else: record a use with decimalai.log_skill_activation() inside a trace.

Use this skill

SKILL.md

npm Downloads to Leads

Take a list of npm packages. Fetch 12 weeks of download data. Compute breakout velocity. Enrich maintainer profiles. Output a ranked lead brief per breakout package with contact signals and an outreach message.

Critical rule: Every package download figure in the output must come from the npm API response. Every maintainer GitHub handle or Twitter username must come from the GitHub API response -- not guessed from the npm username. If the GitHub API did not return a twitter_username field, write "not found on GitHub" -- do not invent one.

Common Mistakes

The agent will want to...	Why that's wrong
Fetch GitHub profiles for every package in the list	Rate limit is 60 req/hr without a token. Enriching steady or declining packages wastes the budget before reaching breakout ones. Only fetch profiles for breakout and watching packages.
Rank packages by raw weekly downloads	Raw downloads favor React and lodash, which are not leads. A package going from 1K to 8K/week is more actionable than React at 50M/week. Velocity score is the signal.
Skip URL-encoding for scoped packages	@org/pkg without encoding causes a 404 from the npm API. Encode @ as %40 and / as %2F for every scoped package name.
Stop the skill when the GitHub rate limit is hit	Degrade gracefully. Present the velocity leaderboard from npm data, skip remaining GitHub enrichments, and add a flag to data_quality_flags. Do not abort.
Write outreach messages without naming the specific package	Generic "I saw your project" messages go unanswered. Every outreach message must name the package, its growth numbers, and a specific connection to the context the user provided.
Include packages below 500 weekly downloads as leads	Below 500/week is noise. The maintainer has no meaningful audience yet. Flag as "too early" but do not present as a lead.

Step 1: Setup Check

echo "GITHUB_TOKEN: ${GITHUB_TOKEN:-not set, unauthenticated rate limit applies (60 req/hr -- enough for ~10 packages)}"

If GITHUB_TOKEN is not set: Continue. Inform the user: "GITHUB_TOKEN is not set. GitHub enrichment is limited to ~10 packages before hitting the rate limit. Add a token at github.com/settings/tokens (no scopes needed)."

No required keys. The npm API and npm registry are fully public with no authentication.

Step 2: Gather Input

Collect from the conversation:

One or more npm package names (unscoped like esbuild, or scoped like @hono/hono)
Optional: a short product context string (used to personalize outreach messages)

If the user gives an npmjs.com URL, extract just the package name. Preserve the full scoped name including @ and org prefix -- encoding is handled in Step 3.

If no packages are provided: Ask: "Which npm packages would you like to analyze? Provide your own, competitors, or a mix. Example: esbuild, @hono/hono, zod, valibot"

python3 << 'PYEOF'
import json, sys

packages_raw = "PACKAGES_HERE"  # comma or newline separated
product_context = "CONTEXT_HERE"  # optional, can be empty string

packages = [p.strip() for p in packages_raw.replace("\n", ",").split(",") if p.strip()]
if not packages:
    print("ERROR: No packages provided.")
    sys.exit(1)

print(f"Packages to analyze: {len(packages)}")
for p in packages:
    print(f"  {p}")

with open("/tmp/npl-input.json", "w") as f:
    json.dump({"packages": packages, "product_context": product_context}, f)
PYEOF

Step 3: Fetch 12-Week Download Data

Use the standalone script if available -- it handles Steps 3, 4, and 5 in one call so you do not need to run the inline code blocks below.

# Check if the script exists
ls scripts/fetch.py 2>/dev/null && echo "script available" || echo "script not found"

If the script is available, run it directly and skip to Step 6:

python3 scripts/fetch.py PACKAGES_HERE --context "CONTEXT_HERE" --output /tmp/npl-script-out.json

Then load the output into the enriched format Step 6 expects:

python3 << 'PYEOF'
import json

out = json.load(open("/tmp/npl-script-out.json"))
# Script output has results array -- split into scored and enriched for Steps 6-8
enriched = [r for r in out["results"] if "profile" in r]
scored = out["results"]
json.dump(scored, open("/tmp/npl-scored.json", "w"), indent=2)
json.dump(enriched, open("/tmp/npl-enriched.json", "w"), indent=2)
json.dump({"packages": [r["package"] for r in scored], "product_context": out.get("product_context", "")},
          open("/tmp/npl-input.json", "w"), indent=2)
print(f"Loaded {len(scored)} packages | {out['breakout_count']} breakout | {out['watching_count']} watching")
PYEOF

If the script is not available, run the inline code below.

Fetch daily download data for each package from the npm Downloads API. Aggregate to weekly buckets.

python3 << 'PYEOF'
import json, urllib.request, sys, time
from datetime import datetime, timedelta, timezone
from collections import defaultdict
import urllib.parse

data = json.load(open("/tmp/npl-input.json"))
packages = data["packages"]

end_date = datetime.now(tz=timezone.utc)
start_date = end_date - timedelta(weeks=13)  # extra week buffer for partial weeks
start_str = start_date.strftime("%Y-%m-%d")
end_str = end_date.strftime("%Y-%m-%d")

results = []
failed = []

for pkg in packages:
    # URL-encode scoped packages: @ -> %40, / -> %2F
    encoded = pkg.replace("@", "%40").replace("/", "%2F")
    url = f"https://api.npmjs.org/downloads/range/{start_str}:{end_str}/{encoded}"

    try:
        req = urllib.request.Request(url, headers={"User-Agent": "npm-downloads-to-leads/1.0"})
        with urllib.request.urlopen(req, timeout=20) as resp:
            raw = json.loads(resp.read())

        # Aggregate daily to weekly by ISO week
        weekly = defaultdict(int)
        for entry in raw.get("downloads", []):
            day = datetime.strptime(entry["day"], "%Y-%m-%d")
            week_key = day.isocalendar()[:2]  # (year, week_num)
            weekly[week_key] += entry["downloads"]

        weeks = [v for k, v in sorted(weekly.items())]
        # Take last 12 complete weekly buckets
        weeks = weeks[-12:]

        results.append({
            "package": pkg,
            "weeks": weeks,
            "total_weeks": len(weeks),
            "current_weekly": weeks[-1] if weeks else 0,
            "status": "ok"
        })
        print(f"  {pkg}: {len(weeks)} weeks, latest week {weeks[-1]:,} downloads")

    except urllib.error.HTTPError as e:
        if e.code == 404:
            failed.append(pkg)
            results.append({"package": pkg, "weeks": [], "total_weeks": 0, "current_weekly": 0, "status": "not_found"})
            print(f"  {pkg}: NOT FOUND (404) -- will be skipped")
        else:
            failed.append(pkg)
            results.append({"package": pkg, "weeks": [], "total_weeks": 0, "current_weekly": 0, "status": f"error_{e.code}"})
            print(f"  {pkg}: HTTP {e.code} error")
    except Exception as e:
        failed.append(pkg)
        results.append({"package": pkg, "weeks": [], "total_weeks": 0, "current_weekly": 0, "status": f"error"})
        print(f"  {pkg}: fetch failed ({e})")

    time.sleep(0.2)  # gentle rate limiting

json.dump(results, open("/tmp/npl-download-data.json", "w"), indent=2)
print(f"\nFetch complete. OK: {len(results) - len(failed)} | Failed/Not found: {len(failed)}")
if failed:
    print(f"Skipped: {', '.join(failed)}")
PYEOF

If all packages return 404 or errors: Stop. Tell the user: "No download data could be fetched. Check that the package names are correct and exist on npmjs.com. Scoped packages must include the full name: @org/package."

Step 4: Compute Velocity Scores

No API call. Pure Python. Compute velocity score, growth ratio, and classify each package.

python3 << 'PYEOF'
import json

raw_results = json.load(open("/tmp/npl-download-data.json"))
scored = []

for item in raw_results:
    pkg = item["package"]
    weeks = item["weeks"]
    status = item["status"]

    if status != "ok" or len(weeks) < 4:
        scored.append({**item, "velocity_score": 0, "growth_pct": 0, "tier": "insufficient_data",
                       "recent_4_avg": 0, "prior_4_avg": 0})
        continue

    recent_4 = sum(weeks[-4:]) / 4
    prior_4 = sum(weeks[-8:-4]) / max(len(weeks) - 4, 1) if len(weeks) >= 8 else sum(weeks[:4]) / max(len(weeks[:4]), 1)
    recent_2 = sum(weeks[-2:]) / 2
    mid_2 = sum(weeks[-4:-2]) / 2 if len(weeks) >= 4 else recent_2

    growth_ratio = recent_4 / max(prior_4, 1)
    acceleration = recent_2 / max(mid_2, 1)
    growth_pct = round((growth_ratio - 1) * 100, 1)

    # Sweet spot multiplier: 500-500K weekly downloads
    if recent_4 < 500:
        noise_factor = max(recent_4 / 500, 0.1)
    elif recent_4 > 500_000:
        noise_factor = max(500_000 / recent_4, 0.1)
    else:
        noise_factor = 1.0

    velocity_score = round(growth_ratio * acceleration * noise_factor * 100, 1)

    # Classify
    if velocity_score > 80 and 500 < recent_4 < 500_000 and growth_ratio >= 1.5:
        tier = "breakout"
    elif velocity_score > 40 and recent_4 >= 500 and growth_ratio >= 1.2:
        tier = "watching"
    elif recent_4 < 500:
        tier = "too_early"
    elif recent_4 >= 500_000:
        tier = "established"
    else:
        tier = "steady"

    scored.append({
        **item,
        "velocity_score": velocity_score,
        "growth_pct": growth_pct,
        "recent_4_avg": round(recent_4),
        "prior_4_avg": round(prior_4),
        "tier": tier
    })

# Sort by velocity_score descending
scored.sort(key=lambda x: x["velocity_score"], reverse=True)

json.dump(scored, open("/tmp/npl-scored.json", "w"), indent=2)

breakout = [p for p in scored if p["tier"] == "breakout"]
watching = [p for p in scored if p["tier"] == "watching"]
too_early = [p for p in scored if p["tier"] == "too_early"]

print(f"Velocity scoring complete:")
print(f"  BREAKOUT: {len(breakout)}")
print(f"  WATCHING: {len(watching)}")
print(f"  STEADY/ESTABLISHED: {len([p for p in scored if p['tier'] in ('steady','established')])}")
print(f"  TOO EARLY (<500/week): {len(too_early)}")
print()
for p in scored[:10]:
    print(f"  {p['tier'].upper():12} {p['package']:30} score={p['velocity_score']:6.1f}  "
          f"{p['recent_4_avg']:>8,}/wk  growth={p['growth_pct']:+.0f}%")

# Stop if nothing worth analyzing
if not breakout and not watching:
    all_too_early = all(p["tier"] in ("too_early", "insufficient_data") for p in scored)
    if all_too_early:
        print("\nERROR: All packages are below the 500 weekly downloads threshold for reliable velocity analysis.")
        print("Try packages with more community adoption.")
        import sys; sys.exit(1)
PYEOF

If all packages are below 500/week: Stop with the message above.

Step 5: Fetch Maintainer Profiles

Only for breakout and watching packages. Fetch npm registry metadata, then GitHub user profiles.

python3 << 'PYEOF'
import json, urllib.request, re, os, time

scored = json.load(open("/tmp/npl-scored.json"))
token = os.environ.get("GITHUB_TOKEN", "")

gh_headers = {"Accept": "application/vnd.github+json", "User-Agent": "npm-downloads-to-leads/1.0"}
if token:
    gh_headers["Authorization"] = f"Bearer {token}"

target_packages = [p for p in scored if p["tier"] in ("breakout", "watching")]
print(f"Fetching profiles for {len(target_packages)} packages (breakout + watching)...")

gh_rate_remaining = 999
enriched = []

for item in target_packages:
    pkg = item["package"]
    profile = {"package": pkg, "npm_maintainers": [], "description": "", "keywords": [],
               "github_owner": None, "github_repo": None, "github_users": [], "npm_homepage": ""}

    # --- npm registry ---
    encoded = pkg.replace("@", "%40").replace("/", "%2F")
    reg_url = f"https://registry.npmjs.org/{encoded}"
    try:
        req = urllib.request.Request(reg_url, headers={"User-Agent": "npm-downloads-to-leads/1.0"})
        with urllib.request.urlopen(req, timeout=20) as resp:
            reg = json.loads(resp.read())

        profile["description"] = reg.get("description", "")
        profile["keywords"] = (reg.get("keywords") or [])[:6]
        profile["npm_homepage"] = reg.get("homepage", "")
        profile["npm_maintainers"] = [m.get("name", "") for m in reg.get("maintainers", []) if m.get("name")]

        # Extract GitHub owner from repository URL
        repo_field = reg.get("repository") or {}
        if isinstance(repo_field, dict):
            repo_url = repo_field.get("url", "")
        else:
            repo_url = str(repo_field)
        gh_match = re.search(r"github\.com[/:]([^/]+)/([^/.]+)", repo_url)
        if gh_match:
            profile["github_owner"] = gh_match.group(1)
            profile["github_repo"] = gh_match.group(2).rstrip(".git")

        print(f"  {pkg}: registry OK | maintainers={profile['npm_maintainers'][:3]} | "
              f"github_owner={profile['github_owner']}")
    except Exception as e:
        print(f"  {pkg}: registry fetch failed ({e})")

    time.sleep(0.1)

    # --- GitHub user profiles ---
    candidates = []
    if profile["github_owner"]:
        candidates.append(profile["github_owner"])
    # Also try npm maintainer usernames (often match GitHub)
    for m in profile["npm_maintainers"][:2]:
        if m and m not in candidates:
            candidates.append(m)

    for username in candidates[:3]:
        if gh_rate_remaining <= 5:
            print(f"  GitHub rate limit low ({gh_rate_remaining} remaining) -- skipping {username}")
            break

        gh_url = f"https://api.github.com/users/{username}"
        req = urllib.request.Request(gh_url, headers=gh_headers)
        try:
            with urllib.request.urlopen(req, timeout=15) as resp:
                gh_rate_remaining = int(resp.headers.get("X-RateLimit-Remaining", 999))
                gh_data = json.loads(resp.read())

            profile["github_users"].append({
                "username": username,
                "name": gh_data.get("name") or username,
                "twitter_username": gh_data.get("twitter_username") or "not found on GitHub",
                "bio": gh_data.get("bio") or "",
                "blog": gh_data.get("blog") or "",
                "company": gh_data.get("company") or "",
                "followers": gh_data.get("followers", 0),
                "public_repos": gh_data.get("public_repos", 0),
                "github_url": gh_data.get("html_url", f"https://github.com/{username}")
            })
            print(f"    GitHub @{username}: {gh_data.get('followers', 0)} followers | "
                  f"twitter={gh_data.get('twitter_username') or 'none'} | rate_remaining={gh_rate_remaining}")
        except urllib.error.HTTPError as e:
            if e.code == 404:
                print(f"    GitHub @{username}: not found")
            else:
                print(f"    GitHub @{username}: HTTP {e.code}")
        except Exception as e:
            print(f"    GitHub @{username}: failed ({e})")

        time.sleep(0.2)

    enriched.append({**item, "profile": profile})

json.dump(enriched, open("/tmp/npl-enriched.json", "w"), indent=2)
json.dump(scored, open("/tmp/npl-scored.json", "w"), indent=2)
print(f"\nEnrichment complete. Profiles fetched: {len(enriched)}")
print(f"GitHub rate limit remaining: {gh_rate_remaining}")
PYEOF

Step 6: Generate Lead Briefs

Print enriched breakout and watching packages, then generate lead briefs and outreach messages.

python3 << 'PYEOF'
import json

enriched = json.load(open("/tmp/npl-enriched.json"))
input_data = json.load(open("/tmp/npl-input.json"))
product_context = input_data.get("product_context", "")

breakout = [p for p in enriched if p["tier"] == "breakout"]
watching = [p for p in enriched if p["tier"] == "watching"]

print("=== DATA FOR LEAD BRIEF GENERATION ===")
print(f"Product context: {product_context or '(none provided)'}")
print()

for item in breakout + watching:
    pkg = item["package"]
    prof = item.get("profile", {})
    gh_users = prof.get("github_users", [])
    primary_gh = gh_users[0] if gh_users else {}

    print(f"PACKAGE: {pkg} ({item['tier'].upper()})")
    print(f"  Velocity score: {item['velocity_score']} | Growth: {item['growth_pct']:+.0f}%")
    print(f"  Recent 4-week avg: {item['recent_4_avg']:,}/week | Prior 4-week avg: {item['prior_4_avg']:,}/week")
    print(f"  Weekly trend (last 8): {item['weeks'][-8:]}")
    print(f"  Description: {prof.get('description', 'none')}")
    print(f"  Keywords: {', '.join(prof.get('keywords', []))}")
    print(f"  npm maintainers: {', '.join(prof.get('npm_maintainers', []))}")
    if primary_gh:
        print(f"  GitHub: @{primary_gh.get('username')} | {primary_gh.get('followers')} followers | "
              f"{primary_gh.get('public_repos')} repos")
        print(f"  Twitter: {primary_gh.get('twitter_username')}")
        print(f"  Bio: {primary_gh.get('bio')}")
        print(f"  Company: {primary_gh.get('company')}")
    else:
        print(f"  GitHub: no profile found")
    print()
PYEOF

Using the package data printed above, generate a lead brief for each BREAKOUT and WATCHING package.

Rules:

Every growth number in the brief must come from the printed data -- do not round or modify
Every GitHub handle and Twitter username must come from the printed data -- write "not found on GitHub" if the field says that
"Why reach out now" must reference the specific growth inflection (weeks, numbers) from the data
"Suggested first message" must name the package and its growth, and if product_context was provided, connect it specifically to that context
No em dashes. No forbidden words: powerful, robust, seamless, innovative, game-changing, streamline, leverage, transform

Write your lead briefs to /tmp/npl-briefs.json with this exact structure:

{
  "lead_briefs": [
    {
      "package": "pkg-name",
      "tier": "breakout",
      "growth_summary": "1-sentence summary of the growth numbers",
      "maintainer_handle": "@github_handle or npm username if no GitHub found",
      "twitter": "@handle or not found on GitHub",
      "github_followers": 0,
      "why_now": "2-3 sentences specific to this package's inflection point",
      "suggested_message": "2-4 sentences. Names the package, the growth, and connects to product_context if provided."
    }
  ]
}

After writing the file, confirm with:

python3 -c "
import json
d = json.load(open('/tmp/npl-briefs.json'))
print(f'Lead briefs generated: {len(d.get(\"lead_briefs\", []))}')
for b in d['lead_briefs']:
    print(f'  {b[\"package\"]} ({b[\"tier\"]}): maintainer={b[\"maintainer_handle\"]}')
"

Step 7: Self-QA

python3 << 'PYEOF'
import json

scored = json.load(open("/tmp/npl-scored.json"))
enriched = json.load(open("/tmp/npl-enriched.json"))
briefs = json.load(open("/tmp/npl-briefs.json"))

failures = []

# Verify: every brief has a real package name from the scored list
real_packages = {p["package"] for p in scored}
for brief in briefs.get("lead_briefs", []):
    if brief.get("package") not in real_packages:
        failures.append(f"Brief for unknown package '{brief.get('package')}' -- removed")

briefs["lead_briefs"] = [b for b in briefs.get("lead_briefs", []) if b.get("package") in real_packages]

# Verify: velocity leaderboard is sorted correctly (checked on scored, not briefs)
sorted_scores = sorted([(p["package"], p["velocity_score"]) for p in scored], key=lambda x: -x[1])
if scored[0]["velocity_score"] < scored[-1]["velocity_score"]:
    failures.append("Scored list not sorted by velocity_score -- re-sorted")
    scored.sort(key=lambda x: x["velocity_score"], reverse=True)

# Verify: no GitHub/Twitter handles in briefs that weren't in GitHub API responses
enriched_gh = {}
for item in enriched:
    for gh_user in item.get("profile", {}).get("github_users", []):
        enriched_gh[gh_user["username"]] = gh_user.get("twitter_username", "not found on GitHub")

for brief in briefs.get("lead_briefs", []):
    twitter = brief.get("twitter", "")
    if twitter and twitter not in ("not found on GitHub", "") and not twitter.startswith("not found"):
        # Verify it came from the API
        found = any(twitter.lstrip("@") == v.lstrip("@") for v in enriched_gh.values() if v != "not found on GitHub")
        if not found:
            failures.append(f"Warning: Twitter handle '{twitter}' for {brief['package']} not verified in GitHub API data")

# Check required fields
for brief in briefs.get("lead_briefs", []):
    for field in ["package", "tier", "growth_summary", "maintainer_handle", "twitter", "why_now", "suggested_message"]:
        if not brief.get(field):
            failures.append(f"Missing field '{field}' in brief for {brief.get('package', '?')}")

# Check for em dashes
briefs_str = json.dumps(briefs)
if "\u2014" in briefs_str:
    briefs_str = briefs_str.replace("\u2014", " - ")
    briefs = json.loads(briefs_str)
    failures.append("Fixed: em dash characters removed from briefs")

# Check for forbidden words
forbidden = ["powerful", "robust", "seamless", "innovative", "game-changing", "streamline", "leverage", "transform"]
full_text = json.dumps(briefs).lower()
for word in forbidden:
    if word in full_text:
        failures.append(f"Warning: forbidden word '{word}' found in briefs -- review before presenting")

output = {
    "scored": scored,
    "enriched": enriched,
    "briefs": briefs,
    "data_quality_flags": failures
}

json.dump(output, open("/tmp/npl-output.json", "w"), indent=2)
print(f"QA complete. Issues: {len(failures)}")
for f in failures:
    print(f"  - {f}")
if not failures:
    print("All QA checks passed.")
PYEOF

Step 8: Save and Present Output

python3 << 'PYEOF'
import json, os
from datetime import datetime, timezone

output = json.load(open("/tmp/npl-output.json"))
scored = output["scored"]
enriched_map = {e["package"]: e for e in output["enriched"]}
briefs_map = {b["package"]: b for b in output["briefs"].get("lead_briefs", [])}
flags = output["data_quality_flags"]
date_str = datetime.now(tz=timezone.utc).strftime("%Y-%m-%d")

breakout = [p for p in scored if p["tier"] == "breakout"]
watching = [p for p in scored if p["tier"] == "watching"]
too_early = [p for p in scored if p["tier"] == "too_early"]
established = [p for p in scored if p["tier"] == "established"]

lines = [
    f"## npm Breakout Report",
    f"Packages analyzed: {len(scored)} | Breakout: {len(breakout)} | Watching: {len(watching)} | Date: {date_str}",
    "",
    "---",
    "",
    "### Velocity Leaderboard",
    "",
    "| Rank | Package | Weekly Downloads | 8-Week Growth | Velocity Score | Status |",
    "|---|---|---|---|---|---|",
]

for i, pkg in enumerate(scored[:15], 1):
    status_label = {"breakout": "BREAKOUT", "watching": "WATCHING", "steady": "steady",
                    "established": "established", "too_early": "too early", "insufficient_data": "no data"}.get(pkg["tier"], pkg["tier"])
    growth_str = f"{pkg['growth_pct']:+.0f}%" if pkg.get("growth_pct") else "n/a"
    lines.append(
        f"| {i} | {pkg['package']} | {pkg['recent_4_avg']:,} | {growth_str} | "
        f"{pkg['velocity_score']} | {status_label} |"
    )

lines += ["", "---", ""]

if breakout or watching:
    lines += ["### Lead Briefs", ""]

    for item in breakout + watching:
        pkg = item["package"]
        brief = briefs_map.get(pkg, {})
        profile = enriched_map.get(pkg, {}).get("profile", {})
        gh_users = profile.get("github_users", [])
        primary_gh = gh_users[0] if gh_users else {}

        lines.append(f"#### {pkg} ({item['tier'].upper()})")
        lines.append(f"Weekly downloads: {item['recent_4_avg']:,}/week (was {item['prior_4_avg']:,} -- {item['growth_pct']:+.0f}% growth over 8 weeks)")
        if profile.get("description"):
            lines.append(f"What it does: {profile['description']}")
        if profile.get("keywords"):
            lines.append(f"Keywords: {', '.join(profile['keywords'])}")
        lines.append("")

        if primary_gh:
            lines.append(f"**Maintainer: @{primary_gh.get('username')}**")
            lines.append(f"- GitHub: {primary_gh.get('followers', 0):,} followers | {primary_gh.get('public_repos', 0)} public repos")
            lines.append(f"- Twitter: {primary_gh.get('twitter_username', 'not found on GitHub')}")
            if primary_gh.get("bio"):
                lines.append(f"- Bio: \"{primary_gh['bio']}\"")
            if primary_gh.get("company"):
                lines.append(f"- Company: {primary_gh['company']}")
            if primary_gh.get("blog"):
                lines.append(f"- Website: {primary_gh['blog']}")
        elif profile.get("npm_maintainers"):
            lines.append(f"**Maintainer (npm only):** {', '.join(profile['npm_maintainers'][:3])}")
            lines.append("- GitHub profile: not found")

        lines.append("")
        if brief.get("why_now"):
            lines.append(f"**Why reach out now:** {brief['why_now']}")
        if brief.get("suggested_message"):
            lines.append(f"\n**Suggested first message:**")
            lines.append(f"> {brief['suggested_message']}")
        lines.append("")
        lines.append("---")
        lines.append("")

if too_early:
    lines += [f"### Too Early ({len(too_early)} packages below 500 weekly downloads)", ""]
    for p in too_early:
        lines.append(f"- {p['package']}: ~{p['recent_4_avg']:,}/week -- revisit when above 500/week")
    lines.append("")

if established:
    lines += [f"### Established Packages (above 500K/week, velocity less meaningful)", ""]
    for p in established:
        lines.append(f"- {p['package']}: ~{p['recent_4_avg']:,}/week")
    lines.append("")

lines += ["---", ""]
lines.append(f"Data quality notes: {'; '.join(flags) if flags else 'None'}")

output_path = f"docs/npm-leads/{date_str}.md"
os.makedirs("docs/npm-leads", exist_ok=True)
open(output_path, "w").write("\n".join(lines))

print("\n".join(lines))
print(f"\nSaved to: {output_path}")
PYEOF

Clean up temp files:

rm -f /tmp/npl-input.json /tmp/npl-download-data.json /tmp/npl-scored.json \
      /tmp/npl-enriched.json /tmp/npl-briefs.json /tmp/npl-output.json

When it activates

Automatically, when a task matches. Use when asked to find evangelists before they are famous, track competitor package momentum, identify breakout npm packages

About this skill

Automatic

Sales & GTM pack

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

varnan-tech / opendirectory

skills/npm-downloads-to-leads/SKILL.md

View on GitHub ↗

License MITImported 10/17/2026Version v1Bundled 3 files · scripts

Illustrative Want to see what varnan-tech/npm-downloads-to-leads would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Related skills

Skills that do similar work, with their headline benchmark lift.

From the same source and nearby in the catalogue.

Ready to supercharge your AI agents?

SkillScore

Safety

What it changes

Track record

Use this skill

npm Downloads to Leads

Common Mistakes

Step 1: Setup Check

Step 2: Gather Input

Step 3: Fetch 12-Week Download Data

Step 4: Compute Velocity Scores

Step 5: Fetch Maintainer Profiles

Step 6: Generate Lead Briefs

Step 7: Self-QA

Step 8: Save and Present Output

When it activates

SkillSafety

Source

Related skills

More in the registry