Przejdź do treści
Biblioteka skilli
DevOps i chmura · Testowanie

verification-before-completion

/verification-before-completion

Sprawdza efekt wdrożenia przez prawdziwą ścieżkę użytkownika, zanim zamknie zgłoszenie lub uzna zmianę za zrobioną.

Licencja
MIT
Ostatnia zmiana repozytorium
12.11.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
2
Skrypty
0
Komentarze HTML
0
Adresy URL w SKILL.md
2
Przestarzałość

Brak sygnałów. Repozytorium zmienione 12.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co robi i kiedy

Sprawdza efekt wdrożenia przez prawdziwą ścieżkę użytkownika, zanim zamknie zgłoszenie lub uzna zmianę za zrobioną.

Wyzwalacze z opisu

is it done?can I close the ticket?Flux/CI/pods are green, are we good?
Oryginalny opis (EN)

Use when about to claim work is done, or asked "is it done?", "can I close the ticket?", "Flux/CI/pods are green, are we good?" after a deploy, merge or config change — prove the effect through the real user path before any success claim.

Jak działa

  1. Overview
  2. The Iron Law
  3. The Gate Function
  4. Common Failures
  5. Verification by Task Type
  6. Red Flags — STOP
  7. Rationalization Prevention
  8. Key Patterns
  9. When To Apply
  10. DevOps depth
Sekcje z nagłówków SKILL.md, w oryginale · 149 linii, 761 słów
Podgląd SKILL.md
---
name: verification-before-completion
description: Use when about to claim work is done, or asked "is it done?", "can I close the ticket?", "Flux/CI/pods are green, are we good?" after a deploy, merge or config change — prove the effect through the real user path before any success claim.
---

# Verification Before Completion

## Overview

**Core principle:** Evidence before claims, always.

**Violating the letter of this rule is violating the spirit of this rule.**

## The Iron Law

```
NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE
```

If you haven't run the verification command in this message, you cannot claim it passes.

## The Gate Function

```
BEFORE claiming any status or expressing satisfaction:

1. IDENTIFY: What command proves this claim?
2. RUN: Execute the FULL command (fresh, complete)
3. READ: Full output, check exit code, count failures
4. VERIFY: Does output confirm the claim?
   - If NO: State actual status with evidence
   - If YES: State claim WITH evidence
5. ONLY THEN: Make the claim

Skip any step = lying, not verifying
```

## Common Failures

| Claim | Requires | Not Sufficient |
|-------|----------|----------------|
| Tests pass | Test command output: 0 failures | Previous run, "should pass" |
| Deploy succeeded | New image digest running + the real endpoint answering through the user's path (`references/devops-proof-table.md`) | `kubectl apply` exit 0, Pod `Running` |
| No errors since the deploy | Requests on the changed path counted in that window (> 0), then its errors (`references/devops-proof-table.md`) | An empty error panel or log search |
| Infra applied correctly | `terraform show` matches intent | `terraform apply` exit 0 |
| Bug fixed | Original symptom tested: passes | Code changed, assumed fixed |
| Pipeline fixed | Green checkmark in CI | Config change committed |
| Agent completed | VCS diff shows changes | Agent reports "success" |
| Requirements met | Line-by-line checklist | Tests passing |

## Verification by Task Type

**Infrastructure apply:**
```bash
terraform show          # confirm state matches intent
# test the actual resource in cloud console/CLI
```

**Container deploy:**
```bash
kubectl --context <ctx> -n <ns> get pods                      # all Running and Ready, restarts not climbing
kubectl --context <ctx> -n <ns> logs <pod> -c <container> --since=10m   # no errors
curl /health            # health check passes
```

**CI/CD change:**
- Trigger the pipeline and watch it pass
- Don't claim "fixed" until the green checkmark exists

**DNS change:**
```bash
dig +trace example.com           # authoritative answer: what is published
dig @<resolver> example.com      # what the users' resolver serves (TTL column = remaining cache)
curl https://example.com     # site loads correctly
```

**Proxy/network change:**
```bash
curl -v https://example.com
tail -n 50 /var/log/nginx/error.log  # no new errors
```

## Red Flags — STOP

- Using "should", "probably", "seems to"
- Expressing satisfaction before verification ("Done!", "Fixed!", "Deployed!", etc.)
- About to commit/push/PR without verification
- Trusting agent success reports
- Relying on partial verification
- Thinking "just this once"
- **ANY wording implying success without having run verification**

## Rationalization Prevention

| Excuse | Reality |
|--------|---------|
| "Should work now" | RUN the verification |
| "I'm confident" | Confidence ≠ evidence |
| "Just this once" | No exceptions |
| "The command exited 0" | Exit code ≠ correct behavior |
| "Agent said success" | Verify independently |
| "The change was small" | Small changes break things too |
| "Partial check is enough" | Partial proves nothing |
| "Different words so rule doesn't apply" | Spirit over letter |
| "Flux/CI says Ready, so they can close the ticket" | Ready = applied. Name the effect check for the ticket's outcome (proof table) |

## Key Patterns

**Infra:**
```
✅ [Run terraform show] [Confirm resource matches intent] "Infrastructure applied correctly"
❌ "Apply exited 0, should be good"
```

**Deploy:**
```
✅ [new digest in containerStatuses[].imageID] [real endpoint via the user's path: 200 + expected body] "Deploy successful"
❌ "Kubectl apply succeeded"
```

**Requirements:**
```
✅ Re-read plan → Create checklist → Verify each → Report gaps or completion
❌ "Tests pass, done"
```

**Agent delegation:**
```
✅ Agent reports success → Check VCS diff → Verify changes → Report actual state
❌ Trust agent report
```

## When To Apply

**ALWAYS before:**
- ANY variation of success/completion claims
- ANY expression of satisfaction
- ANY positive statement about work state
- Committing, PR creation, task completion
- Moving to next task
- Delegating to agents
- Answering "is it done?" / "can I close the ticket?" about someone else's change: give the effect check, not a yes

**Rule applies to:**
- Exact phrases
- Paraphrases and synonyms
- Implications of success
- ANY communication suggesting completion/correctness

## DevOps depth

For infrastructure work, also apply `references/devops-proof-table.md` — what each green signal proves, and the effect check that proves the rest.
Pliki w paczce (2)
  • SKILL.md 5 KB
  • references/devops-proof-table.md 23 KB

Wersje

RepozytoriumWynikLicencja
Ta strona · nikospavlopoulos/devops-mutant-powers100MIT
jjuraszek/pi-gauntlet100MIT
GanyuanRan/Aegis100MIT
mvschwarz/openrig100Apache-2.0
FrancoStino/opencode-skills-collection100MIT
foryourhealth111-pixel/Vibe-Skills100Apache-2.0
rootcastleco/rei-skills100MIT
Montinou/triqual100MIT
coco-research/coco100MIT

Migawka bazy z 13.11.2026 · wersja paczki 2fadb53c21a0 (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 0 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

  • 100spot-verification-before-completion
    Wymaga dowodów, zanim praca zostanie uznana za zakończoną, działającą lub gotową do wdrożenia.
  • 100sr-verification-before-completion
    Wymusza uruchomienie poleceń weryfikacyjnych i potwierdzenie wyniku, zanim praca zostanie uznana za ukończoną lub naprawioną.
  • 100deploy
    Prowadzi przepływ wdrożenia po scaleniu PR: feature flagi, wdrożenie canary lub blue-green, weryfikacja smoke testami i wyzwalacze wycofania, aby bezpiecznie wypuścić zmianę na produkcję.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.