personalize-repo_skill

Overview

This skill tailors a bootstrapped repository into a project-specific operating system by capturing intent, constraints, and preferences, then updating governance docs and producing a repo-local REPO_PROFILE.json. It is interactive: it asks a short set of focused questions before making any edits, and it only runs when explicitly invoked. The goal is a minimal, deterministic contract for downstream automation and clear, aligned docs.

How this skill works

The skill scans the repo root to detect existing top-level docs, an existing REPO_PROFILE.json, and simple stack hints (package manifests, Makefile, CI). It runs a bounded interview to collect project intent and tradeoffs, proposes precise doc changes, then updates docs in place and creates or updates REPO_PROFILE.json following conservative rules. All edits are minimal, explicit, and documented in a short completion report.

When to use it

• When converting a generic scaffold into a project-specific governance and execution contract
• Before running automated planning or code-generation skills that rely on repository assumptions
• When onboarding a new team or clarifying operational boundaries and quality expectations
• If the repo lacks a canonical REPO_PROFILE.json or the existing profile needs deliberate updates
• Only when you are prepared to answer a short set of interview questions

Best practices

• Answer the 5–8 interview questions up front; if you say 'defaults', the skill will proceed with conservative assumptions and list them
• Do not expect deep code analysis—this is configuration and doc personalization, not implementation
• Avoid silent rewrites: review proposed deltas before the skill writes files
• Use explicit 'none' or 'unknown' in the profile rather than guessing commands
• Keep edits minimal and explicit; if a policy or invariant is added, record it in the completion report

Example use cases

• Onboarding: capture a 30-day definition of done and lock down non-negotiable safety checks
• Stabilization: move from a generic scaffold to a repo-specific profile so CI and agents behave deterministically
• Audit prep: make testing, linting, and release commands explicit in the repo profile
• Decision capture: translate interview answers into targeted updates to architecture, calibration, and roadmap docs