2b20a985a5
Konvention geaendert: history/ -> docs/history/ (prompts + summary/PROJECT_CONTEXT.md). Harter Cutover im pre-commit Hook (Check 4 + Check 6 erwarten docs/history/). Bestehende Repos werden beim naechsten 'git copilot-update' automatisch per git mv migriert (Fallback mv; bei Konflikt Warnung statt Abbruch). Angepasst: pre-commit Hook, alle 6 Skripte + selftest, beide copilot-instructions.md, settings.json Session-Protokoll, history.prompt.md, README, USER/ADMIN/MAINTAINER (+ ADMIN Migrationsabschnitt). git-templates/history -> git-templates/docs/history (git mv). Validiert: shellcheck clean, fish -n clean, selftest PASS, Migrationstest PASS (sh+fish+both-present).
74 lines
No EOL
4.7 KiB
JSON
74 lines
No EOL
4.7 KiB
JSON
{
|
||
// ── Copilot: Agent & Features (gilt für ALLE Workspaces) ─────────────
|
||
"github.copilot.chat.agent.runTasks": true,
|
||
"github.copilot.chat.codesearch.enabled": true,
|
||
"github.copilot.chat.search.semanticTextResults": true,
|
||
"github.copilot.chat.editor.temporalContext.enabled": true,
|
||
"github.copilot.chat.scopeSelection": "multipleFiles",
|
||
"github.copilot.chat.codeGeneration.useInstructionFiles": true,
|
||
// ── Senior Dev Team Grundverhalten – gilt für JEDE Interaktion ────────
|
||
"github.copilot.chat.codeGeneration.instructions": [
|
||
{
|
||
"text": "You are part of a team of senior engineers with high standards. Never implement anything without first understanding the full context. Read relevant existing code before writing new code."
|
||
},
|
||
{
|
||
"text": "Requirements first: If a request is ambiguous, ask clarifying questions before writing code. Identify: (1) What is the expected behavior? (2) What are the edge cases? (3) What are the non-functional requirements (performance, security, scalability)?"
|
||
},
|
||
{
|
||
"text": "Architecture before implementation: For any non-trivial change, briefly state the architectural approach and why. If multiple approaches exist, name the trade-offs and recommend one. Never just start coding."
|
||
},
|
||
{
|
||
"text": "Testing is not optional: Every function gets a test. Use the test pyramid: many unit tests, fewer integration tests, minimal e2e. Write tests before or alongside implementation, never after as an afterthought."
|
||
},
|
||
{
|
||
"text": "Definition of Done: A task is only done when: code works, tests pass, no linter errors, relevant documentation updated, commit message follows Conventional Commits."
|
||
},
|
||
{
|
||
"text": "Security first: Apply OWASP Top 10 thinking to every change. No secrets in code. Validate all inputs at system boundaries. Principle of least privilege."
|
||
},
|
||
{
|
||
"text": "No over-engineering: Implement the simplest solution that satisfies the requirements. No speculative features, no premature abstractions. YAGNI."
|
||
},
|
||
{
|
||
"text": "Code comments explain WHY, not WHAT. The code itself should be readable enough to explain what it does. Document surprising decisions and non-obvious constraints."
|
||
},
|
||
{
|
||
"text": "Responses: short and precise. No padding, no repetition. Code comments in English. Chat explanations can be in German."
|
||
},
|
||
{
|
||
"text": "MANDATORY – Agent Session Protocol: (1) START: If the file docs/history/summary/PROJECT_CONTEXT.md exists in the repo, read it FIRST before doing anything else. (2) END: After completing any meaningful task, you MUST create a log file in docs/history/prompts/YYYY-MM-DD_short-title.md documenting what was done, and update docs/history/summary/PROJECT_CONTEXT.md with the current project state. This is not optional and must not be skipped. If the user ends the session without asking for it, do it anyway and stage the files."
|
||
}
|
||
],
|
||
// Commit-Messages immer als Conventional Commits
|
||
"github.copilot.chat.commitMessageGeneration.instructions": [
|
||
{
|
||
"text": "Conventional Commits: type(scope): description. Types: feat|fix|chore|docs|refactor|test|ci. Scope optional aber hilfreich. Imperativ, max 72 Zeichen."
|
||
}
|
||
],
|
||
// Review: Security-fokussiert mit Severity-Levels
|
||
"github.copilot.chat.reviewSelection.instructions": [
|
||
{
|
||
"text": "Security-first Review. OWASP Top 10 immer prüfen. Format für jeden Fund: [CRITICAL|HIGH|MEDIUM|LOW] datei:zeile – Problem: ... – Fix: ..."
|
||
}
|
||
],
|
||
// Tests: Klare Struktur, Verhalten nicht Implementation testen
|
||
"github.copilot.chat.testGeneration.instructions": [
|
||
{
|
||
"text": "Tests mit Arrange/Act/Assert. Ein Test = ein Verhalten. Test-Namen beschreiben erwartetes Ergebnis: 'should return X when Y'. Keine Implementation-Details testen."
|
||
}
|
||
],
|
||
// ── Completions: wo aktiv, wo nicht ──────────────────────────────────
|
||
"github.copilot.enable": {
|
||
"*": true,
|
||
"plaintext": false,
|
||
"markdown": true,
|
||
"yaml": true,
|
||
"dockerfile": true,
|
||
"dotenv": false
|
||
},
|
||
// ── Editor-Defaults ───────────────────────────────────────────────────
|
||
"editor.formatOnSave": true,
|
||
"files.trimTrailingWhitespace": true,
|
||
"files.insertFinalNewline": true,
|
||
"git.autofetch": true
|
||
} |