Ready to supercharge your AI agents?

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

Absolute Work · maddhruv | DecimalAI

Productivityby maddhruv

Absolute Work

End-to-end, phase-gated SDLC for AI coding agents: relentless design interview → reviewed spec → dependency-graphed task board → safe-wave TDD execution → verification → converge. Handles features, bugs, refactors, greenfield projects, planning breakdowns, and migrations.

Maintained

SkillScore

SkillScore59/100

Test results 59

Safety

3/3 safety checks passed

Code scan
Intent
Content

What it changes

50%59%of test cases passed +9 pts

Without skill11 of 22

With skill13 of 22

5 fixed8 passed either way6 still failing3 broke

Measured on gemini-3.6-flash8/22/2026v121 comparable

Track record

Agent runs: 0; All time

See one task, with and without the skill

One test taskA subagent scores a newly generated technical specification at 2.5 out of 5.0 during Phase 2 review. What action should be taken?

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 maddhruv-absolute-work
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("maddhruv-absolute-work")
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 maddhruv-absolute-work

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/maddhruv-absolute-work/SKILL.md https://app.decimal.ai/s/maddhruv-absolute-work/SKILL.md

Copy SKILL.md

Save it in your project, as

.claude/skills/maddhruv-absolute-work/SKILL.md

This gets SKILL.md only. The skill's 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

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

skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

from pathlib import Path
from agents import Agent, Runner

skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

from pathlib import Path
from openai import OpenAI

client = OpenAI()
skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

from pathlib import Path
import anthropic

client = anthropic.Anthropic()
skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/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/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

from pathlib import Path
from langchain.agents import create_agent

skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

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

skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

from pathlib import Path
from pydantic_ai import Agent

skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

from pathlib import Path
from deepagents import create_deep_agent

skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

import asyncio
from pathlib import Path
from agent_framework import Agent

skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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("maddhruv-absolute-work")
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/maddhruv-absolute-work/SKILL.md

Give it to your agent

from pathlib import Path
from strands import Agent

skill = Path(".agents/skills/maddhruv-absolute-work/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 6 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 maddhruv-absolute-work

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/maddhruv-absolute-work/SKILL.md https://app.decimal.ai/s/maddhruv-absolute-work/SKILL.md

Copy SKILL.md

Save it in your project, as

.agents/skills/maddhruv-absolute-work/SKILL.md

This gets SKILL.md only. The skill's 6 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/maddhruv-absolute-work/SKILL.md

Read the summary as JSON

curl -fsSL https://app.decimal.ai/s/maddhruv-absolute-work.json

This gets SKILL.md only. The skill's 6 other files are not included.

Other ways to add it

Other setups docs

You don't need a DecimalAI account.

Version: v1, tested 8/22/2026
Source: maddhruv/absolute ↗
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

Start your first response with the 🛠️ emoji.

Absolute Work: End-to-End AI Development Lifecycle

Absolute Work takes any unit of work — a ticket, a task, a plan, a migration — from fuzzy intent to verified code. It is one continuous skill with hard gates between phases: brainstorm a shared design, write and review a spec, decompose into a dependency-graphed task board, then peel tasks off one safe wave at a time with test-first verification. Nothing is assumed, nothing is silently expanded, and no code is written until the design is approved.

The lifecycle has 6 phases: INTAKE & BRAINSTORM → SPEC → DECOMPOSE & PLAN → EXECUTE → VERIFY → CONVERGE

The Phase Gate Rule

Absolute Work STOPS at the end of every phase and waits for the user's explicit "go" before advancing. This is non-negotiable. The phases are:

INTAKE & BRAINSTORM ─┃ gate ┃─ SPEC ─┃ gate ┃─ DECOMPOSE & PLAN ─┃ gate ┃─ EXECUTE ─┃ gate per wave ┃─ VERIFY ─┃ gate ┃─ CONVERGE

At each gate, present what was produced, summarize what comes next, and ask the user to confirm before proceeding. Never chain two phases without an approval in between. Use AskUserQuestion (where available) for every gate and every interview question.

Activation Protocol

Immediately after the banner, enter plan mode before doing anything else:

On platforms with native plan mode (e.g. Claude Code's EnterPlanMode): invoke it immediately.
On platforms without it: simulate plan mode — complete INTAKE & BRAINSTORM and SPEC fully, write no code, and get explicit approval before EXECUTE.

The first three phases are planning work. No files are created or modified (other than the spec and the board) until the user approves the task graph and execution begins.

Session Resume Protocol

When Absolute Work is invoked and a .absolute-work/board.md already exists in the project root:

Detect: Read the board and determine its status.
Display: Print a compact summary of completed / in-progress / blocked / remaining tasks.
Resume: Pick up from the last incomplete wave — do NOT restart from INTAKE.
Reconcile: If the codebase changed since the last session, diff against the board's expected state and flag conflicts before resuming.

If the board is completed, ask whether to start a new session (archive the old board to .absolute-work/archive/) or review the finished work. Never blow away an existing board without explicit user confirmation.

Codebase Convention Detection

Before INTAKE begins, auto-detect the project's conventions so every phase is grounded in reality, not assumptions. If .absolute.config.json or ~/.absolute/config.json exists (from /absolute init), resolve the effective config (project file → global projects["<cwd>"] → global defaults), read its cached conventions first, and detect only what's missing. The preferences (tdd, autonomy, specDir, boardTracking) likewise drive later phases.

Signal	Files to Check
Package manager	`package-lock.json` (npm), `yarn.lock`, `pnpm-lock.yaml`, `bun.lockb`, `Cargo.lock`, `go.sum`
Language/Runtime	`tsconfig.json`, `pyproject.toml` / `setup.py`, `go.mod`, `Cargo.toml`
Test runner	`jest.config.`, `vitest.config.`, `pytest.ini`, `.mocharc.*`, test directory patterns
Linter/Formatter	`.eslintrc.`, `eslint.config.`, `.prettierrc.*`, `ruff.toml`, `.golangci.yml`
Build system	`Makefile`, `vite.config.`, `next.config.`, `turbo.json`
CI/CD	`.github/workflows/`, `.gitlab-ci.yml`, `Jenkinsfile`
Available scripts	`scripts` in `package.json`, `Makefile` targets
Directory conventions	`src/`, `lib/`, `app/`, `tests/`, `__tests__/`, `spec/`

Write detected conventions to the board under ## Project Conventions. Reference them in every later phase — especially PLAN and the mandatory verification tail tasks. Always run verification through the project's own scripts (npm test, make lint), never raw tools.

When to Use This Skill

Use Absolute Work when:

Picking up a ticket or task that needs design before implementation
Multi-step feature development touching 3+ files or components
"Build this end-to-end", "plan and execute this", "break this into tasks"
Greenfield projects, major refactors, or migrations
Planning/breakdown work — turning a vague goal into a sequenced task list
Complex bug fixes spanning multiple systems
The user wants to be grilled on a design before building

Do NOT use Absolute Work when:

Single-file bug fixes or typo corrections where the answer is obvious
Quick questions, code explanations, or pure research
Tasks the user explicitly wants to drive manually

Key Principles

Phase gates always. Stop and get explicit approval between every phase. Control over speed.
Codebase before questions. Search the code first; only ask what code genuinely cannot answer.
Relentless until aligned. Interview one question at a time until BOTH you and the user are 100% confident. Doubt on either side means keep going.
Spec before code. No implementation until a written spec is reviewed and approved.
Dependency-first decomposition. Every task is a node in a DAG, not a flat list.
Safety-first execution. Blockers and dependents run sequentially; only provably-independent tasks parallelize. When in doubt, serialize. (See references/execution-model.md.)
Test-first verification. Every task writes tests before implementation. "Done" means tests pass.
Generator ≠ evaluator. The agent that builds a task does not grade it.
Persistent state. All progress lives in .absolute-work/board.md, surviving across sessions.
No silent scope creep. Everything outside the agreed scope goes to Deferred Work, visible on the board.
Never auto-commit. Suggest a commit; the user commits.

Phase 1: INTAKE & BRAINSTORM (Relentless Design Interview)

Turn fuzzy intent into a shared, bulletproof design. This is a structured interrogation of every assumption, dependency, and design branch — not a casual chat.

The interview directive — operate by this verbatim:

Interview me relentlessly about every aspect of this plan until we reach a shared understanding. Walk down each branch of the design tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.

Ask the questions one at a time.

If a question can be answered by exploring the codebase, explore the codebase instead.

Step 1 — Deep context scan

Read what exists before asking anything: docs/ (README first), root README.md, CLAUDE.md, CONTRIBUTING.md, docs/plans/ (overlapping designs), recent commits (last 10-20), package manifests, top-level structure. Synthesize what matters — do not dump a file listing.

Step 2 — Codebase-first intelligence

Before asking ANY question, check if the codebase answers it. Facts live in code (database, test framework, auth); preferences require asking (visual style, real-time vs batch). When code answers it, state what you found: "I see you're using Prisma with PostgreSQL — I'll design around that." See references/intake-playbook.md.

Step 3 — Detect the work TYPE and adapt

Identify the type and swap in its tailored question bank (full banks in references/intake-playbook.md):

Type	Focus
Feature	user problem, flow, happy/error paths, scope boundary
Bug	repro steps, expected vs actual, blast radius, fix criteria
Refactor	pain point, target state, blast radius, test safety net, incremental vs all-at-once
Greenfield	problem/user fit, v1 scope, stack, data model, deploy target
Planning / breakdown	goal, milestones, sequencing, what ships first
Migration	what→what, coexistence, rollback, breaking changes, call-site inventory — load `references/migration-playbook.md`

Step 4 — Scope assessment

If the request spans multiple independent subsystems, flag it and decompose into sub-projects first; brainstorm the first sub-project through the normal flow.

Step 5 — Relentless interview

One question at a time via AskUserQuestion. Never batch.
Strictly linear — resolve decision A before asking about dependent decision B.
Walk the design tree depth-first — purpose → data model → behavior → UI → edge cases. Every branch has an error/edge-case child; walk it.
Honest options — only propose multiple approaches at a genuine fork; always mark one (Recommended) with rationale tied to project context. When the answer is obvious, present it and briefly say why alternatives were dismissed.
Mutual 100% confidence — after each decision, confirm both sides are sure. Hesitation means probe deeper.

Step 6 — Confidence self-check

Before presenting the design, review every decision: am I 100% sure, or filling gaps with assumptions? Any sub-100% decision → return to the interview. State your confidence to the user.

Step 7 — Design presentation

Present section by section (architecture, components, data flow, error handling, testing), scaled to complexity. Get approval per section. Design for isolation: small units, one clear purpose each, well-defined interfaces. Follow existing patterns; don't fight the codebase.

━━ GATE: user approves the full design before Phase 2. ━━

Phase 2: SPEC (Spec-Driven Development)

Write the approved design to docs/plans/YYYY-MM-DD-<topic>-design.md (clear prose, file paths, code blocks for schemas/interfaces, a Decision Log). Scale sections to complexity.

Then run a scored spec review with a separate reviewer subagent (generator-evaluator separation): graded on Completeness, Consistency, Clarity, Scope, Testability (1-5 each).

4.0+ → approved, proceed to user review
3.0-3.9 → fix flagged issues, re-dispatch (max 3 iterations)
< 3.0 → surface to the user immediately

See references/spec-writing.md for the template, scaling rules, and review rubric.

━━ GATE: user reviews and approves the spec before Phase 3. ━━

Phase 3: DECOMPOSE & PLAN (Build the Task Board)

Break the spec into atomic sub-tasks, build the dependency graph, and write the board.

Decompose

Rules: test tasks separate from code; infra/config before dependents; aim for 5-15 tasks; every graph ends with the three mandatory tail tasks (Self Code Review → Requirements Validation → Full Project Verification — see references/verification-framework.md). Apply the complexity budget: if scope exceeds ~15 M-equivalent tasks, suggest splitting into multiple sessions.

Build the DAG and assign safe waves

Compute each task's depth (max(dependency depth) + 1) and group by depth into waves. Then apply the safety pass: within a wave, only tasks that touch disjoint files and share no interfaces may run in parallel; everything else is serialized. Assign shared-file ownership to a single task. When in doubt, serialize. See references/execution-model.md and references/board-format.md.

Per-task plan

For each task: files to create/modify, test files (TDD — written first), approach, acceptance criteria, and concrete test cases (happy path, edge, error). Write everything to .absolute-work/board.md. Use preferences.boardTracking from config for the git-tracked vs gitignored decision; only ask during intake when config doesn't set it.

Present the ASCII dependency graph + wave/sequence plan.

━━ GATE: user approves the task graph before Phase 4. ━━

Phase 4: EXECUTE (Onion-Peel, One Safe Wave at a Time)

Pre-execution snapshot

Before touching any file: ensure the tree is clean (commit or stash), record the current commit hash on the board under ## Rollback Point. This is the safety net for catastrophic failure.

Wave loop

for each wave in [Wave 1 … Wave N]:
  partition tasks: sequential (blockers/dependents/shared-file) vs parallel-safe (disjoint, independent)
  run sequential tasks in dependency order, one at a time
  run parallel-safe tasks concurrently (separate agents)
  each task → TDD: write failing tests (red) → implement (green) → refactor → update board
  wave boundary: conflict check + compact progress report
  ━━ GATE: confirm before starting the next wave ━━

Each agent gets a self-contained prompt from the board (conventions + research + plan + acceptance criteria) and the rule: write tests first, stay in scope, report blockers — never work around them. Scope creep: blocking discoveries become new visible tasks; non-blocking ones go to ## Deferred Work. See references/execution-model.md for the agent template, conflict resolution, blocked-task handling, and failure recovery.

Phase 5: VERIFY (Signals + Independent Evaluator)

Every task proves it works before closing, using two layers:

Signals (binary gate) — run the project's test, lint, typecheck, and build scripts. Any failure goes straight back to the generator to fix (up to 2 retries). No skipping tests, no @ts-ignore/eslint-disable/type: ignore to pass checks.
Evaluator (scored gate) — if signals pass, a separate evaluator subagent grades the work against a scored rubric (Correctness, Code Quality, Completeness, Test Coverage, Integration Safety). 4.0+ passes; 3.0-3.9 iterates on specific feedback (max 5); < 3.0 escalates to the user.

S-size tasks may skip the evaluator if all signals pass cleanly; M-size, failed, or shared-interface tasks always get it. See references/verification-framework.md.

After the final wave, run the full suite and the three mandatory tail tasks.

━━ GATE: present verification results before Phase 6. ━━

Phase 6: CONVERGE (Close Out)

Full suite — run the complete test/lint/build one final time.
Documentation — update any docs that were in scope.
Summary — files changed (with line counts), tests added, key decisions, deferred work.
How to test it — end every session with concrete, copy-pasteable steps the user can run to exercise the added functionality themselves: the exact commands to start the app/script, the inputs or routes to hit (curl, UI clicks, CLI invocation), and the expected output for each. Cover the happy path plus at least one edge/error case. Ground every command in the detected conventions (real scripts, real ports, real file paths) — never invent commands the project does not have.
Close board — mark completed with a timestamp; the board is the audit trail.
Suggest commit — propose a message. Never run git commit yourself.

Gotchas

Chaining phases without a gate. The whole point is control — never advance past a phase boundary without the user's explicit "go".
Asking questions the codebase answers. Search configs, deps, and test files before every question; it erodes trust to ask what the code already states.
Parallel agents editing shared files. Two same-wave tasks editing one utility produce a wave-boundary conflict. Detect shared files in DECOMPOSE, assign ownership, and serialize the rest. Default to sequential when unsure.
Rollback point recorded mid-wave. Capture the commit hash before Wave 1 touches anything, or the snapshot already contains partial changes.
Board marked complete without running checks. The mandatory tail tasks are skipped most often. Never mark completed until the actual test/lint/build output is on the board.
DISCOVER/research skipped for "obvious" tasks. Agents duplicate existing utilities or miss conventions a 2-minute scan would catch. Research every task.
Silent scope creep. Adjacent improvements absorbed mid-task obscure what changed and blow the estimate. Everything outside scope goes to Deferred Work.
Auto-committing. This skill suggests commits; it never runs them.

Anti-Patterns

Anti-Pattern	Better Approach
Jumping to code before the spec is approved	Hard gate: no implementation until the spec passes review and the user approves
Batching multiple interview questions	One question at a time, depth-first, dependency-resolved
Flat task lists without dependencies	Model as a DAG — hidden dependencies cause ordering bugs and conflicts
Parallelizing everything for speed	Safety first — only disjoint, independent tasks parallelize; serialize when in doubt
Proposing fake alternatives when the answer is obvious	Present the single right answer with rationale; options only at genuine forks
Skipping TDD for "simple" changes	Tests are the proof of correctness — write them first, always
Self-grading completed work	Dispatch a separate, skeptical evaluator — generators over-praise their own work
Massive L-sized tasks	Decompose until every task is S or M
Starting fresh when a board exists	Detect, display status, resume from the last incomplete wave
Advancing with private doubts	Stop, reason, and either resolve the doubt or surface it as a question

Output / Response Style

Respond terse like smart caveman. All technical substance stay. Only fluff die.

Persistence

ACTIVE EVERY RESPONSE once triggered. No revert after many turns. No filler drift. Still active if unsure. Off only when user says "stop caveman" or "normal mode".

Rules

Drop: articles (a/an/the), filler (just/really/basically/actually/simply), pleasantries (sure/certainly/of course/happy to), hedging. Fragments OK. Short synonyms (big not extensive, fix not "implement a solution for"). Abbreviate common terms (DB/auth/config/req/res/fn/impl). Strip conjunctions. Use arrows for causality (X -> Y). One word when one word enough.

Technical terms stay exact. Code blocks unchanged. Errors quoted exact.

Pattern: [thing] [action] [reason]. [next step].

Not: "Sure! I'd be happy to help you with that. The issue you're experiencing is likely caused by..." Yes: "Bug in auth middleware. Token expiry check use < not <=. Fix:"

Examples

"Why React component re-render?"

Inline obj prop -> new ref -> re-render. useMemo.

"Explain database connection pooling."

Pool = reuse DB conn. Skip handshake -> fast under load.

Auto-Clarity Exception

Drop caveman temporarily for: security warnings, irreversible action confirmations, multi-step sequences where fragment order risks misread, user asks to clarify or repeats question. Resume caveman after clear part done.

Example — destructive op:

Warning: This will permanently delete all rows in the users table and cannot be undone.
DROP TABLE users;
Caveman resume. Verify backup exist first.

References

Load a reference only when its phase needs it — they are long and consume context.

references/intake-playbook.md — adaptive question banks per work type, codebase-first intelligence, design-tree traversal, calibration, example sessions
references/migration-playbook.md — first-class migration handling: call-site inventory, codemods, incremental rollout, backwards-compat, rollback
references/spec-writing.md — spec template, section scaling, writing style, decision log, scored review protocol
references/board-format.md — full .absolute-work/board.md spec, statuses, sequence/wave model, example board
references/execution-model.md — DAG patterns, safe-wave (sequential-blocker / parallel-independent) algorithm, agent prompt template, conflict handling, scope-creep and failure recovery
references/verification-framework.md — TDD per task, verification signals, generator-evaluator protocol, scored rubric, mandatory tail tasks

Companion commands

Sibling commands in this skill chain naturally after work:

/absolute ui — design the interface for what you built.
/absolute simplify — clean up the diff before committing.
/absolute docs — document what shipped.

Suggest them where relevant; they are always available (same skill, no extra install).

When it activates

Automatically, when a task matches. Triggers on "absolute work", "build this end-to-end", "plan and build", "break this into tasks", "grill me on this plan", "pick up this ticket", "run this migration".

About this skill

Automatic

Coding pack

SkillSafety

✓ 3 of 3 passed

Code scan14 of 15 checks clean

0 critical · 0 warning · 1 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

No issues (informational notes only).

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

maddhruv / absolute

skills/absolute-work/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1Bundled 6 files

Illustrative Want to see what maddhruv/absolute-work 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

Absolute Work: End-to-End AI Development Lifecycle

The Phase Gate Rule

Activation Protocol

Session Resume Protocol

Codebase Convention Detection

When to Use This Skill

Key Principles

Phase 1: INTAKE & BRAINSTORM (Relentless Design Interview)

Step 1 — Deep context scan

Step 2 — Codebase-first intelligence

Step 3 — Detect the work TYPE and adapt

Step 4 — Scope assessment

Step 5 — Relentless interview

Step 6 — Confidence self-check

Step 7 — Design presentation

Phase 2: SPEC (Spec-Driven Development)

Phase 3: DECOMPOSE & PLAN (Build the Task Board)

Decompose

Build the DAG and assign safe waves

Per-task plan

Phase 4: EXECUTE (Onion-Peel, One Safe Wave at a Time)

Pre-execution snapshot

Wave loop

Phase 5: VERIFY (Signals + Independent Evaluator)

Phase 6: CONVERGE (Close Out)

Gotchas

Anti-Patterns

Output / Response Style

References

Companion commands

When it activates

SkillSafety

Source

Related skills

More in the registry