scope-creep-response

Installation
SKILL.md

Scope Creep Response Skill

Scope creep is rarely malicious — it's a client who genuinely can't tell the difference between "small tweak" and "new workstream," talking to a freelancer afraid that any boundary costs the relationship. The damage comes from the two failure modes: absorbing everything silently (resentment, then a blowup over something trivial) or lawyering every request (death by friction). This skill runs the middle path: classify honestly, respond proportionally, and make the structure hold the boundary so the person doesn't have to.

What This Skill Produces

  • The classification — in-scope / gray / clearly-out, argued against the actual agreement text, including the honest case for the client's reading
  • The graduated response — goodwill grant, flag-and-track, change order, or renegotiation — with ready-to-send wording
  • The change-order artifact — small, friendly, priced; a form not a confrontation
  • The prevention layer — scope language, revision counts, and the "great idea — phase 2" ritual for the next contract

Required Inputs

Ask for these if not provided:

  • What the agreement actually says — the scope text verbatim; if scope was never written down, that's the finding, and the response changes (you can't cite what doesn't exist)
  • The asks so far — list them; one gray-zone request and a drip of twelve "tiny things" are different situations
  • Relationship context — client value, history, how earlier extras were handled (silently absorbed extras set precedent that must be un-set gently)
  • Their goal — keep the client happily, get paid for the extras, or exit gracefully — the same classification routes to different responses
Installs
1
GitHub Stars
1.5K
First Seen
Oct 1, 2026