成

The Shipyard

ShipLift

Make your real work visible.

ShipLift analyzes your repository and combines it with short, adaptive check-ins to turn engineering work into evidence-backed achievements, goals, standups, 1:1 talking points, and CV bullets — for engineers whose real impact doesn't fully show up in a commit log or survive six months of memory.

Install ShipLiftView on GitHub

What Is ShipLift?

Building is only half the craft. Being able to show it is the other half.

ShipLift is an AI coding-agent skill, built to the Agent Skills standard, that analyzes a software repository and converts engineering work into meaningful, evidence-based achievements. Its own stated goal isn't to summarize commits — it's to understand what you shipped, why it mattered, and how to communicate that clearly.

It combines two sources of evidence: code evidence from Git, and human evidence from Pulse — a short, adaptive check-in that captures the work Git structurally can't see, like unblocking a teammate or making a hard technical call.

It's built for engineers, not managers or recruiters — people who are good at building things but, by their own admission in the project's philosophy, not always good at remembering or explaining them later. It sits alongside the normal engineering workflow, at the moments where that work needs to become visible: standups, 1:1s, goal check-ins, quarterly reviews, and CV updates.

The Problem

Git can see your commits. It can't see why they mattered.

A commit log shows activity, not impact. "Reviewed 5 PRs" doesn't tell you whether that mattered. Real engineering work — an investigation, a hard call, unblocking a teammate — often leaves no trace in Git at all. By the time a performance conversation or a CV update comes around, most of that context is gone, and what's left tends to drift toward one of two failure modes: understated ("I just fixed some bugs") or overstated (a list of buzzwords with no evidence behind them).

ShipLift's answer is an Evidence Engine that links related work into one story, an Impact Analysis pass that keeps activity and impact separate, and an Anti-BS Engine that refuses to let either failure mode through — nothing invented, nothing understated either.

How It Works

From two evidence sources to one honest story

  1. 01

    Input — Code + Human Evidence

    Git supplies code evidence. Pulse — a short, adaptive Q&A — supplies human evidence for the work Git can't see.

  2. 02

    Evidence Engine

    Linking, Strength, and Signals: related evidence gets connected into one story instead of staying a list of separate activities.

  3. 03

    Impact Analysis

    Separates Activity, Contribution, Impact, and Outcome. If the impact isn't known, ShipLift says "Impact: Unknown" instead of guessing.

  4. 04

    Anti-BS Engine

    Validation pass that refuses to invent percentages, productivity gains, revenue impact, or leadership claims not supported by evidence.

  5. 05

    Output — Career Intelligence

    Achievements, Goals, Standup, 1:1 talking points, and CV bullets — all traceable back to real evidence.

The core idea, in the project's own words: "Make the value of the work clearer, not bigger."

What Does ShipLift Do?

Eight capabilities, all evidence-first

  1. Quarterly Achievements

    Analyzes the current calendar quarter and returns 5–7 strongest achievements, grouping related commits, reviews, and fixes into one engineering story instead of a raw activity list.

  2. Pulse

    A short (30–90 second), adaptive, local-first check-in that captures work Git can't see — helping someone, an investigation, an unblock, a technical decision. "Nothing" and "Not sure" are valid answers; Pulse never pressures you into manufacturing an achievement.

  3. Evidence Linking

    Connects related evidence — e.g. an investigation → root cause → fix → regression tests → shared findings — into one meaningful story instead of five disconnected bullet points.

  4. Impact Analysis

    Keeps Activity, Contribution, Impact, and Outcome distinct. "Reviewed 5 PRs" does not automatically become a productivity percentage.

  5. Anti-BS Engine

    The core rule: never make the engineer sound more impressive than the evidence allows. No invented percentages, revenue impact, cost savings, or leadership claims.

  6. Career Signals

    Identifies recurring patterns over time — e.g. repeated code reviews plus helping teammates plus technical discussions — without automatically treating a pattern as a proven achievement.

  7. Blind Spots

    Reports where evidence is limited instead of asserting a negative — "ShipLift has limited evidence of mentoring this quarter," not "you did not mentor anyone."

  8. Goals

    Connects achievements and evidence to your professional goals, runs a SMART validation pass, and can surface recurring themes as possible goals — never presenting a suggestion as an official goal until you confirm it.

The Workflow

Not another tool to remember — a cadence to run

ShipLift doesn't sit inside a build or deploy pipeline — it runs alongside the normal engineering workflow, at the points where work needs to become visible:

  1. 01

    During the week

    ShipLift Pulse

    Run it whenever you finish something meaningful that might not show up clearly in Git.

  2. 02

    During the quarter

    ShipLift Goals

    Check how your work supports your stated goals.

  3. 03

    Before standup

    ShipLift Standup

    Combines recent Git activity with Pulse evidence into a Done / Next / Blockers update.

  4. 04

    Before your 1:1

    ShipLift 1:1

    Prepares evidence-backed talking points: delivery, impact, challenges, goal progress, ownership.

  5. 05

    End of quarter

    ShipLift Quarter

    Analyzes the full quarter and returns the strongest evidence-backed achievements.

  6. 06

    CV update

    ShipLift CV

    Turns your engineering history into CV-ready bullets, scoped by year, quarter, range, or level if you ask.

Installation

Install ShipLift

Prerequisite

Node.js (for the recommended npx skills installer) or the GitHub CLI, plus one of the four supported agents.

Installation — recommended

The Skills CLI auto-detects the agents on your machine:

npx skills add bishoy-bishai/ShipLift

Configuration — target a specific agent

Claude Code

Supported
npx skills add bishoy-bishai/ShipLift -a claude-code

Or via GitHub CLI

gh skill install bishoy-bishai/ShipLift --agent claude-code --scope user

Codex

Supported
npx skills add bishoy-bishai/ShipLift -a codex

Or via GitHub CLI

gh skill install bishoy-bishai/ShipLift --agent codex --scope user

Cursor

Supported
npx skills add bishoy-bishai/ShipLift -a cursor

Or via GitHub CLI

gh skill install bishoy-bishai/ShipLift --agent cursor --scope user

Antigravity

Supported
npx skills add bishoy-bishai/ShipLift -a antigravity

Or via GitHub CLI

gh skill install bishoy-bishai/ShipLift --agent antigravity --scope user

Install globally, all four agents at once

npx skills add bishoy-bishai/ShipLift \
  -g \
  -a claude-code \
  -a codex \
  -a cursor \
  -a antigravity

Manual installation

Clone the repository and copy SKILL.md, references/, and scripts/ into whichever skills directory your agent reads from:

git clone https://github.com/bishoy-bishai/ShipLift.git

Verification

Open or restart your AI coding agent and run:

ShipLift Pulse

If it's installed correctly, the agent recognizes the command and starts the Pulse flow rather than treating it as plain text.

First usage

ShipLift Quarter

Analyzes your current quarter and returns 5–7 evidence-backed achievements.

To update an installation later: npx skills update (or gh skill update for a GitHub CLI install).

Quick Start

From nothing to your first achievement list

  1. 01

    Install:

    npx skills add bishoy-bishai/ShipLift
  2. 02

    Configure — start capturing evidence Git can't see:

    ShipLift Pulse
  3. 03

    Run the main analysis:

    ShipLift Quarter
  4. 04

    Verify: confirm the achievements returned trace back to real commits, reviews, or Pulse entries — not invented metrics.

Examples

Real usage, from the repository

What you want to do

Summarize what I actually accomplished this quarter

Command

ShipLift Quarter

Result

12 tests added, 4 bugs fixed, 6 components updated, and 8 PRs reviewed become one story: "Test Quality & Regression Protection — increased the automated test suite by 35%, added regression tests for critical flows, improved coverage of important edge cases."

What you want to do

Capture something that won't show up in Git

Command

ShipLift Pulse

Result

A 30–90 second adaptive Q&A — did you help anyone, review code, make a technical decision, unblock the team — stored locally under ~/.shiplift/ and fed into every other command.

What you want to do

Prepare for my manager 1:1

Command

ShipLift 1:1

Result

Evidence-backed talking points covering what was delivered, impact, challenges, goal progress, technical ownership, and collaboration.

What you want to do

Update my CV with real, measured results

Command

ShipLift CV

Result

"Increased the automated test suite by 35% and added regression coverage for critical frontend flows." — never an invented metric. Can be scoped, e.g. ShipLift CV Senior.

Why ShipLift

Writing the code is not the finish line.

Most career systems ask "what did you accomplish?" ShipLift asks a better question, in its own words: "What evidence do we have of the value you created?"

Engineers are usually good at building things. They're not always good at remembering them — and worse at explaining them six months later. ShipLift's writing constitution enforces the same discipline in both directions: no corporate buzzwords dressing up small work as huge, and no understatement burying real work either. The target voice is a senior engineer explaining their work to another engineer.

Technical Details

Architecture

Like the other Arsenal skills, ShipLift is a markdown-and-script skill package — no application to deploy:

ShipLift/
├── SKILL.md
├── README.md
├── VALIDATION.md
├── references/
│   ├── achievement-framework.md
│   ├── commands.md
│   ├── evidence-matrix.md
│   ├── impact-engine.md
│   ├── intelligence-rules.md
│   ├── anti-bs-rules.md
│   ├── metrics.md
│   ├── output-templates.md
│   ├── goals-engine.md
│   ├── career-evidence-engine.md
│   └── pulse-engine.md
└── scripts/
    ├── git-analysis.sh
    ├── pulse-store.sh
    └── pulse_store.py

Storage

Pulse is local-first: entries are stored under ~/.shiplift/ and feed Quarter, Goals, Standup, 1:1, and CV.

Roadmap — potential future evidence sources

The architecture is designed so new evidence sources can feed the same Evidence Engine. Documented as potential — not yet shipped: GitHub, Linear, Jira, Slack, Notion, Calendar.

License: MIT — source on GitHub.

Compatibility

Supported AI Coding Agents

  • Claude CodeSupported
  • CodexSupported
  • CursorSupported
  • AntigravitySupported

ShipLift follows the Agent Skills standard, installed via npx skills or gh skill install. It runs wherever your agent and Node.js or Git are available — the repository doesn't document OS-specific requirements beyond that.

FAQ

Frequently asked questions

What is ShipLift?
ShipLift is an AI coding-agent skill that analyzes your repository — combined with short, adaptive human check-ins called Pulse — to convert engineering work into evidence-backed achievements, goals, standups, 1:1 talking points, and CV-ready bullets.
What does ShipLift do?
It runs an Evidence Engine over Git activity and Pulse check-ins, links related evidence into one story, separates activity from actual impact, and passes everything through an Anti-BS Engine that refuses to invent percentages, revenue impact, or leadership claims not supported by evidence.
Who should use ShipLift?
Engineers who want an accurate, evidence-based record of their work for standups, 1:1s, performance conversations, and CV updates — instead of trying to remember or reconstruct it months later.
How do I install ShipLift?
Run npx skills add bishoy-bishai/ShipLift (optionally targeting a specific agent with -a), or gh skill install bishoy-bishai/ShipLift --agent <name>. Manual installation by cloning the repo and copying SKILL.md, references/, and scripts/ into your agent's skills directory is also documented.
How does ShipLift fit into a developer workflow?
It runs alongside the coding workflow, not inside a build/deploy pipeline: Pulse during the week for evidence Git can't see, then Standup, 1:1, Goals, Quarter, and CV at their natural cadence points.
Is ShipLift open source?
Yes — MIT License, source on GitHub.
Can ShipLift work with Claude Code?
Yes — install with npx skills add bishoy-bishai/ShipLift -a claude-code, or the equivalent gh skill install command.
Can ShipLift work with Codex?
Yes — install with npx skills add bishoy-bishai/ShipLift -a codex.
Can ShipLift work with Cursor?
Yes — install with npx skills add bishoy-bishai/ShipLift -a cursor.
Can ShipLift work with Antigravity?
Yes — install with npx skills add bishoy-bishai/ShipLift -a antigravity.

Built In The Open

Explore the source code, documentation, examples, and development history.

View ShipLift on GitHub

The Arsenal

Four tools. One Master.

The Shipyard

ShipLift

Make your real work visible.

Install ShipLiftView on GitHub

Part of the SIMURAI Arsenal