generated: '2026-08-25'
method: searched
source: https://github.com/LedgerHQ/agent-skills
note: >-
  PROVIDER-PUBLISHED, not generated. All four skills below are Ledger's own, saved verbatim from
  raw.githubusercontent.com/LedgerHQ/agent-skills/main on 2026-08-25. Ledger distributes them with
  `npx skills add ledgerhq/agent-skills`, and since wallet-cli 2.1.0 the wallet-cli skill also ships
  embedded in the binary (`wallet-cli skill install --agent cursor|claude|codex|agents`, with
  `wallet-cli skill doctor` to detect drift between an installed copy and the binary's). The
  Ledger Enterprise CLI ships a fifth bundled skill (`ledger-enterprise-cli skill install`) which is
  NOT in the public agent-skills repository — it is emitted by the binary — so it is recorded in
  cli/ledger-cli.yml rather than saved here.
repository: https://github.com/LedgerHQ/agent-skills
install: npx skills add ledgerhq/agent-skills
license: none declared in the repository
maturity: >-
  Ledger labels both agent surfaces early development - "Breaking changes may happen between
  releases."
skills:
  - file: ledger-dmk-implementation.md
    name: ledger-dmk-implementation
    upstream: https://github.com/LedgerHQ/agent-skills/tree/main/skills/dmk/ledger-dmk-implementation
    raw: https://raw.githubusercontent.com/LedgerHQ/agent-skills/main/skills/dmk/ledger-dmk-implementation/SKILL.md
    role: primary
    api: Ledger Device Management Kit
    description: >-
      Execute DMK signing operations safely — initialise the SDK, establish a device session, verify
      device state, open the correct chain app, perform the signing or device-management operation,
      return the result. Five-step execution process with human-in-the-loop gates, error
      classification and timeout bounds.
    refs:
      - dmk-sdk-reference.md
      - dmk-code-patterns.md
      - dmk-platform-patterns.md
    connectors: [dmk-intent-vocabulary, dmk-business-logic]
  - file: ledger-dmk-intent-vocabulary.md
    name: dmk-intent-vocabulary
    upstream: https://github.com/LedgerHQ/agent-skills/tree/main/skills/dmk/dmk-intent-vocabulary
    raw: https://raw.githubusercontent.com/LedgerHQ/agent-skills/main/skills/dmk/dmk-intent-vocabulary/SKILL.md
    role: connector
    api: Ledger Device Management Kit
    description: >-
      Map informal developer intent ("sign a tx", "find my Ledger", "is this device real?") onto the
      right DMK component, operation and API method before any implementation begins.
  - file: ledger-dmk-business-logic.md
    name: dmk-business-logic
    upstream: https://github.com/LedgerHQ/agent-skills/tree/main/skills/dmk/dmk-business-logic
    raw: https://raw.githubusercontent.com/LedgerHQ/agent-skills/main/skills/dmk/dmk-business-logic/SKILL.md
    role: connector
    api: Ledger Device Management Kit
    description: >-
      Explain the domain model and design rationale behind DMK — Clear Signing vs blind signing,
      Secure Channel, Ledger OS (BOLOS), Device Actions vs Commands, sessions, transports,
      derivation paths and Genuine Check.
  - file: ledger-wallet-cli-usage.md
    name: wallet-cli-usage
    upstream: https://github.com/LedgerHQ/agent-skills/tree/main/skills/wallet-cli/wallet-cli-usage
    raw: https://raw.githubusercontent.com/LedgerHQ/agent-skills/main/skills/wallet-cli/wallet-cli-usage/SKILL.md
    role: primary
    api: Ledger Wallet CLI
    cli: '@ledgerhq/wallet-cli'
    description: >-
      Drive the Ledger Wallet CLI from natural language — account discover, receive, balances,
      operations, send, swap quote/execute/status, genuine-check, asset lookup, and Ledger Key Ring
      (LKRP) encryption — and map informal requests to the right command.
    refs:
      - https://raw.githubusercontent.com/LedgerHQ/agent-skills/main/skills/wallet-cli/wallet-cli-usage/references/business-logic.md
agent_config_note: >-
  Ledger's llms.txt and several docs pages carry machine-readable `agent_skills:` frontmatter with
  primary/refs roles and raw URLs, and instruct a reading agent to fetch and apply them as working
  context. That is an unusually deliberate agent-onboarding convention and is the reason these
  skills are `searched` rather than generated.