verify

द्वारा factory-ai

ड्रॉइड-नियंत्रण वर्कफ़्लो के लिए पृष्ठभूमि ज्ञान -- सीधे आह्वान नहीं किया जाता। प्रतिबद्धताओं के विरुद्ध वितरणीय सत्यापन।

npx skills add https://github.com/factory-ai/factory-plugins --skill verify

Verify

The orchestrator routed you here. This atom checks the final deliverable against the commitments made at the start of the workflow.

Inputs

You receive:

  1. Commitments from the command's parse step -- the promises made about what the deliverable would contain
  2. Compose outputs -- the finished artifact(s) and their metadata

Video deliverables

Technical checks

Run ffprobe on the final .mp4:

ffprobe -v quiet -print_format json -show_format -show_streams <video>
CheckPass condition
Exists and playsffprobe exits 0, duration > 0
Resolution1920x1080 (or target resolution)
Pixel formatyuv420p
File sizeUnder 5 MB for GitHub embeds (25 MB hard limit)
DurationMust fall within the compose pacing table's target range for this demo type: 30-45s (single feature), 45-75s (side-by-side comparison), 60-120s (multi-phase). Below the minimum is a failure — re-compose with a lower speed factor or re-capture with more steps.
FilenameIncludes PR number or meaningful identifier

Commitment checks

Walk through each commitment from the parse step:

CommitmentHow to verify
Title cardVideo starts with a static frame showing PR info (check first 5s)
Side-by-side layoutVideo shows two panels with a divider
Showcase polishResolution is 1920x1080, window chrome and rounded corners visible
Keystroke overlayPill overlays appear at interaction points
EffectsEffects matching the committed tier are present (utilitarian: zoom/keystroke; full: spotlight, zoom, callout, keystroke)
Speed noteTitle card mentions playback speed

Content checks

  • Every claim from the "what to prove" analysis has visible evidence in the video
  • Both states (before/after, input/result) appear on screen
  • No dead time longer than 3 seconds without visible activity

Screenshot/snapshot deliverables

For proofs

CheckPass condition
Evidence existsScreenshots/snapshots at every claimed proof point
Environment statedDriver, terminal/browser, OS identified
Conclusion presentEvidence explicitly supports or refutes the claim
Before/after pairedIf comparison, both branches shown at same capture points

For QA reports

CheckPass condition
Step coverageEvery defined test step has a result (PASS/FAIL)
Evidence attachedScreenshots/snapshots at every step
Failures documentedFailed steps have evidence and description
Report structuredMarkdown report follows the QA template

Failure handling

If any check fails:

  1. Identify which stage produced the problem (capture or compose)
  2. Report the specific failure: "Side-by-side layout was committed but the output is a single panel"
  3. Go back to the failed stage and fix it
  4. Re-verify after the fix

Do not report a deliverable as complete until every commitment is met.

Output

## Verification

### Technical
- Resolution: 1920x1080 ✓
- Duration: 42s ✓
- Size: 3.2 MB ✓
- Format: yuv420p ✓

### Commitments
- [x] Title card with PR info
- [x] Side-by-side comparison layout
- [x] Showcase hero preset applied
- [x] Keystroke overlay visible

### Content
- [x] Fork creates independent session (visible at 0:18)
- [x] History diverges after fork (visible at 0:32)

All commitments met. Deliverable ready.

factory-ai की और Skills

wiki
factory-ai
एक रिपॉजिटरी पढ़ें, फिर परस्पर जुड़े दस्तावेज़ीकरण पृष्ठों का एक सेट तैयार करें जो बताता है कि कोड क्या करता है और यह कैसे एक साथ फिट होता है। आउटपुट droid-wiki/ निर्देशिका में मार्कडाउन फ़ाइलों का एक सेट है, जिसे droid wiki-upload के माध्यम से Factory पर अपलोड किया जाता है।
official
threat-model-generation
factory-ai
किसी रिपॉजिटरी के लिए STRIDE-आधारित सुरक्षा खतरा मॉडल तैयार करें। सुरक्षा निगरानी स्थापित करते समय, आर्किटेक्चर बदलावों के बाद, या सुरक्षा ऑडिट के लिए उपयोग करें।
official
agent-repellent
factory-ai
अपने कोडबेस की AI-सहायता प्राप्त विकास उपकरणों के प्रति प्रतिरोधक क्षमता का आकलन करें।
official
blame
factory-ai
गिट इतिहास जांच से पैटर्न और विसंगतियां उजागर होती हैं।
official
cobol-converter
factory-ai
COBOL माइग्रेशन योजना अनुवादित आउटपुट के साथ।
official
dating-profile
factory-ai
प्रौद्योगिकी स्टैक और आर्किटेक्चर पर आधारित संगतता प्रोफ़ाइल।
official
feng-shui
factory-ai
आपके कोडबेस के लिए वास्तुकला सामंजस्य मूल्यांकन।
official
obituary
factory-ai
अप्रचलित और जीवन-अंत कोड पर पूर्वव्यापी दृष्टि।
official