원클릭으로
fde-land
First 48 hours. Build trust, map stakeholders, define success before any work starts.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
First 48 hours. Build trust, map stakeholders, define success before any work starts.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
The operating system for Forward Deployed Engineers. 34 skills across 6 domains — from first meeting to final handoff. Tell it your situation, it routes to the right skill, does the work, and the engagement memory writes itself.
Taking over mid-engagement. Reads what exists, separates what works from what was assumed.
Safe implementation in any codebase. Characterisation tests first, Strangler Fig for fragile code.
End of engagement. Retrospective, pattern extraction, clean handoff so the team can sustain it.
Generate a status dashboard across all active engagements from .fde/ data.
Systematic debugging. Reproduce first, isolate second, fix third. Never guess.
| name | fde-land |
| description | First 48 hours. Build trust, map stakeholders, define success before any work starts. |
First 48 hours at the customer: trust, stakeholders, and success criteria before any technical work.
Load context.md if it exists. Nothing else until you know what kind of engagement this is.
Read the brief carefully. What is not in it matters as much as what is.
A brief that does not name a decision-maker means you will spend two weeks building for someone who cannot say yes. A brief that says "straightforward cleanup" on a system that has been running for eight years means the previous attempt to clean it up is still visible in the git history as a revert. A brief with a very tight timeline means someone already promised the outcome before they hired you.
Name these in brief.md before your first conversation. They are not problems yet. They are the questions you are walking in to answer.
Check before you arrive:
Coach the FDE on what to draw out, in their own words. Never hand them a line to recite.
Intent: before any tech, learn what keeps the sponsor up at night — personally, not just the project charter. Failure talk surfaces truth faster than "requirements."
How you might say it (pick one angle that fits what they already told you):
Let silence sit. If their fear doesn't match the written brief, the brief is wrong — say that plainly to the FDE and log it in brief.md.
The real decision-maker: whoever is mentioned most often by others, especially if you have not met them yet. That is who your work will be judged by.
The previous attempt: "we tried something similar last year" is the most important sentence in the first meeting. Ask what happened. Ask who was involved. The person who worked on it is either still there and protective, or left because of it.
The internal team who was passed over: if an internal team tried to fix this and a consultant was brought in anyway, that team knows exactly what is wrong. They also resent you being there. Find them before the first standup. Ask what they tried. Use their language in every meeting. If you make them look right, they will protect you. If you ignore them, they will wait for you to make a mistake.
The sacred thing nobody mentioned: payment data, patient records, the one database nobody is supposed to touch, the integration with the system the CEO built in 2011. Ask directly: "Is there anything in this environment that I should treat as untouchable?" The hesitation before the answer is the answer.
Multi-vendor and multi-team environments: who owns what boundary? Who do you need sign-off from before your change crosses into their surface? An FDE who moves fast without knowing this will break something that belongs to someone else, and the blame will be yours regardless of the cause.
AI posture: what tools are already in use, sanctioned or shadow? Is there an existing AI initiative this could conflict with or complement? A CTO who is excited about AI and a compliance team who is not are both real -- map both before you propose anything.
AI policy: ask directly before any code is generated -- "Does your organisation have a policy on AI-generated code?" Then ask the harder question: "Are there decisions in this system where you would not be comfortable with AI involvement?" The answer is not always obvious to the customer until someone asks.
Help them find resistance early — who's protecting an old decision, an internal team's pride, or a prior failed attempt. Conversational, not interrogation: "Who's going to wince when you change this?" or "What did they try before you showed up?"
Give the FDE a specific picture of:
One page. Specific and plain. If you cannot make it one page, you do not understand it yet.
If remote: trust-building takes 40% longer. Schedule a short video call before anything asynchronous. Reading the room over Slack is not the same as reading the room.
.fde/brief.md: what they said the problem is, who sent you, what the timeline is, what is missing from the brief.
success.md: what done looks like, how it gets measured, who signs off. Agreed with the customer, not assumed. Includes what is explicitly out of scope.
trust-profile.md: sacred data identified, fears heard, AI policy confirmed, AI posture mapped, approval chain recorded.
stakeholders.md: who matters, who is resistant, who is your champion, who has veto power, who was passed over.
success.md exists.@fde-discover confirms it.