Skip to main content
evidence.json contains a list of items that agents can use to verify your skills. Each item points to something real and public — a GitHub repository, a certification, a portfolio piece, or a publication. Agents can fetch and inspect these links to confirm your capabilities rather than relying solely on your self-reported profile.json.
Evidence URLs are checked for reachability during scoutica validate. A URL that returns a non-200 status will produce a validation warning.

​
Schema

​
schema_version
string
required
Protocol version this evidence registry conforms to. Current value: "0.1.0".
​
items
object[]
required
Array of evidence items.

​
Evidence types

TypeDescription
github_repoA public GitHub repository. Agents can inspect language distribution, commit history, and last activity.
websiteA personal or project website.
portfolioA portfolio page or design showcase.
certificateAn industry certification with a verifiable certificate URL.
articleA blog post or technical article.
reviewA third-party review or testimonial.
referenceA reference from a colleague or employer.
photoA photo demonstrating work (e.g., physical project, installation).
videoA video demonstrating work or a talk.
case_studyA detailed write-up of a project or engagement.
publicationA research paper, academic publication, or book.
otherAny other form of verifiable public evidence.

​
Trust levels

Trust levels describe how strongly an evidence item has been independently verified. Higher levels carry more weight with employer agents.
LevelLabelMeaning
0Self-assertedYou added this URL yourself. No external verification has occurred.
1URL reachableThe Scoutica validator confirmed the URL returns a 200 response.
2Content verifiedAn agent has inspected the content and confirmed it matches the claimed skills.
3Peer-endorsedA trusted third party has attested to this item.
4Blockchain-verifiedThe item is anchored to an immutable public ledger (roadmap feature).
Trust levels 0 and 1 are assigned by the CLI validator. Trust levels 3 and 4 require external processes not yet automated — they are part of the roadmap.

​
Example

Based on the sample card in the repository: 
{
  "schema_version": "0.1.0",
  "items": [
    {
      "type": "github_repo",
      "title": "CloudDeploy — Kubernetes Deployment Toolkit",
      "url": "https://github.com/alexchen/cloud-deploy",
      "description": "Open-source Kubernetes deployment toolkit with GitOps workflow and automated rollbacks.",
      "skills_demonstrated": ["Kubernetes", "Go", "CI/CD Pipelines", "Infrastructure as Code"]
    },
    {
      "type": "github_repo",
      "title": "FastAPI Starter — Production API Template",
      "url": "https://github.com/alexchen/fastapi-starter",
      "description": "Production-ready FastAPI template with async database, auth, and observability built in.",
      "skills_demonstrated": ["Python", "FastAPI", "PostgreSQL", "Docker"]
    },
    {
      "type": "certificate",
      "title": "AWS Solutions Architect Associate",
      "url": "https://www.credly.com/badges/example-badge-id",
      "description": "Validates expertise in designing distributed systems on AWS.",
      "skills_demonstrated": ["AWS", "Cloud Infrastructure"]
    }
  ]
}

​
How agents verify evidence

For github_repo items, agents can inspect the repository directly:
  • Does the repository exist and is it publicly accessible?
  • What programming languages are present?
  • How many commits does it have and when was it last active?
  • Does the candidate appear as a contributor?
For certificate items, agents check the provided URL to confirm the credential is live and valid. For article and publication items, agents treat them as demonstrations of domain knowledge and writing ability.