Skip to main content
The Scoutica Protocol is used across the full recruiting lifecycle — by candidates creating discoverable profiles, companies building recruiting agents, platforms adding card import, and teams publishing composite cards.

​
Quick reference

If you are…Start here
A professional creating your cardScenario 1: Candidate creates their card
A company building a recruiting agentScenario 2: AI recruiter finds candidates
A platform adding Scoutica Protocol importScenario 3: Job board auto-import
Building autonomous agent negotiationScenario 4: Agent-to-agent negotiation
Implementing evidence verificationScenario 5: Evidence verification
Creating team or org cardsScenario 6: Team card
Exploring blockchain verificationScenario 7: Blockchain-verified card
Maria, a senior backend engineer, wants to make her skills discoverable by AI agents without sharing her data with job boards.Flow:
  1. Maria runs scoutica scan ~/maria-cv/ --with gemini
  2. The CLI reads her CV, certifications, and portfolio documents
  3. Documents are piped to her local Gemini CLI — nothing leaves her machine
  4. The AI generates profile.json, rules.yaml, evidence.json, and SKILL.md
  5. Maria runs scoutica validate — 5 checks pass
  6. Maria runs scoutica publish — the CLI commits and pushes to GitHub
scoutica scan ~/maria-cv/ --with gemini
scoutica validate
scoutica publish
Result: Maria’s Skill Card is published at github.com/maria/skills. Her data never left her laptop. Any AI agent can now discover and evaluate her profile.For the full creation walkthrough, see the Create Your Skill Card guide.
TechCorp’s AI agent needs to fill a Senior DevOps Engineer position. It searches the Scoutica Protocol registry.Flow:
  1. The agent queries the registry: skills=["Kubernetes","Terraform"], seniority="senior"
  2. The registry returns a list of matching card URLs
  3. The agent fetches profile.json and rules.yaml for each candidate in parallel
  4. Each card is scored and checked against the job requirements
  5. Candidates whose rules.yaml auto-rejects the offer (e.g., salary below minimum) are removed automatically
  6. The agent contacts only the candidates who passed all checks
Example scoring results:
CandidateScoreOutcome
Maria85Contacted
Alex72Contacted (note: requires on-site option)
Sam91Auto-rejected — salary below minimum
Key insight: Sam was never contacted. The agent respected the rules.yaml auto-reject for salary. No wasted time for either party.
DevJobs.io lets applicants import their Scoutica Protocol card instead of filling out application forms.Flow:
  1. The candidate clicks “Import Scoutica Card” and pastes their card URL
  2. The platform fetches scoutica.json to get the card URL and basic details
  3. The platform fetches profile.json and auto-fills name, title, skills, and experience
  4. The platform fetches rules.yaml and runs a pre-screen against the job’s salary and remote policy
  5. If there is a mismatch (e.g., role pays 75K, candidate minimum is 85K), the platform warns the candidate before they apply
Benefit: Candidates don’t fill forms. The platform pre-screens and surfaces mismatches upfront, saving time for both sides.
Building a platform integration? See the developer guide for the API spec and code examples.
Maria’s personal AI agent receives an inbound opportunity and negotiates autonomously on her behalf.Flow:
  1. TechCorp’s hiring agent sends: {role: "Lead DevOps", salary: 95K, remote: hybrid}
  2. Maria’s agent loads rules.yaml and checks all conditions:
    • Salary 95K > minimum 85K — pass
    • Remote hybrid — pass (policy: flexible)
    • Industry fintech — pass (not in blocked list)
  3. Maria’s agent responds with a counter-offer: {salary: 110K, equity: "1%"}
  4. TechCorp’s agent counters: {salary: 105K, equity: "0.5%", signing_bonus: 5K}
  5. Maria’s agent accepts — the total package is within acceptable range
  6. Maria receives a notification: “Interview scheduled: Lead DevOps at FinTech Corp, 105K + equity”
Result: Three negotiation rounds completed in seconds. Maria is only notified when there is a concrete interview to schedule.
Before an interview, an employer’s agent verifies Maria’s claimed skills against her evidence.json.Flow:
  1. The agent loads evidence.json (5 items)
  2. For each GitHub repo: fetches stars, languages, and commit counts via the GitHub API
  3. For each certification URL: sends a HEAD request to verify the link is live
  4. For each portfolio link: checks HTTP status
Example results:
Evidence itemResult
github.com/maria/infra-toolkitVerified — Go + Python match claimed skills
Terraform certification URLVerified — 200 OK
github.com/maria/k8s-operatorVerified — Kubernetes expertise confirmed
Portfolio siteVerified — live
Old certification linkFailed — 404 Not Found
Trust score: 4/5 verified (80%). The agent proceeds — strong evidence with one minor dead link.
Keep your evidence.json up to date. Dead certification links reduce your trust score even if the underlying skill is real.
SquadAlpha — a team of 4 engineers — publishes a composite team card that companies can hire as a unit.Structure:The team card aggregates data from individual member cards:
  • Combined skills across all members
  • Team rules — availability, engagement model, team rate
  • Collective evidence — joint projects and publications
Members:
MemberRoleEntity type
MariaLead DevOpshuman
AlexSenior Backendhuman
SamML Engineerhuman
CodeBotAI Agentai_agent
This enables companies to hire entire teams rather than assembling individuals one by one. The team card uses the same three-zone privacy model as individual cards.
In a future release, cards can be verified on-chain using Soulbound Tokens (SBTs) on the Base L2 network.Flow:
  1. The candidate mints an SBT with a hash of their profile
  2. Verifiers (e.g., former employers, certification bodies) endorse specific skills on-chain
  3. Employers query the smart contract to check endorsements
  4. The trust level is elevated to “Blockchain-verified” — a 2x multiplier on the trust score
Why SBTs? Soulbound Tokens are non-transferable — they prove that an endorsement was given to a specific person and cannot be sold or transferred.
Blockchain verification is on the roadmap. Today’s V1 uses URL-based evidence verification (Scenario 5) as the trust mechanism.