Why sealed

Open protocol, not a private format

The container, digests, mint attestation, and TrustView semantics come from the open .skill protocol. This registry implements product UX on top: catalog, activity, Skill Score, and sealed delivery. No tokens, NFTs, or paid protocol tiers.

Skill Score v2 (WIP)

Skill Score is computed from protocol scoreSignals (sealed, digest, permissions, source, real Rekor only when present, release profile, registry activity, not revoked). It is not a malware scan and never claims "safe to run."

See Score + seal proof on TrustView

Reproducible trust dashboard

Public numbers on installs, publishes, revocations, and top capabilities come from the append-only log. Verify any digest with inclusion proofs against the signed tree head. No private counters required.

/api/v1/log/dashboard/api/log/tree-head/log

Verify a package

# Digest or ref (registry VerifyResult)
curl -s -X POST https://skillerr.com/api/verify \
  -H 'content-type: application/json' \
  -d '{"ref":"skillerr/find-skills"}'

# Inclusion proof
curl -s 'https://skillerr.com/api/log/proof?digest=sha256:…'

# Wire catalog row
curl -s https://skillerr.com/api/skills/skillerr/find-skills

Public links

• Open .skill protocol

  Spec, trust ladder, container layout, host requirements

• Protocol reference

  Profiles, integrity digests, mint vs continuity, TrustView

• Verify API

  POST a .skill (proxied to docs.skillerr.com/api/verify)

• Submit a skill

  Send a SKILL.md URL for manual review and sealing

• llms.txt

  Machine-readable catalog pointers for agents

• Transparency log

  Append-only publishes and installs you can audit

• Reference implementation

  CLI + @skillerr/core / protocol (open source)

• Open log dashboard

  Installs, publishes, revokes, top capabilities from the public Merkle log

• Signed tree head

  GET SignedTreeHead for independent verification