agentleFS
StripeKnown publisherstripe/link-cli/skills/use-link-identity/SKILL.md

use-link-identity

Skill896 starsChanged 30 days agoMITView on GitHub

Use Link CLI's identity commands when a service requests a Link Agent Attestation Token (AAT) or signed user claims such as email. Covers acquiring and managing attestations, obtaining identity credentials, and presenting selected claims to a verifier through an HTTP client or browser automation.

  • Installs packages

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use Link Agent Wallet v0.27.0 or later.

    line 26
  2. Very specific:

    Only request needs Link authentication.

    line 29
  3. Very specific:

    Check errors and attestations.

    line 46
  4. Very specific:

    Use a new file outside ~/.link-cli/identity/attestations.

    line 67

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Link identity

These commands are in beta and appear in link-cli identity --help without a preview flag. Use them when a service accepts Link attestations or signed user claims.

Service requestsCommandJSON field → HTTP header
A Link Agent Attestation Token (AAT), often via WWW-Authenticate: PrivateTokenidentity attestations popauthorization → Authorization
Signed claims, often via WWW-Authenticate: Identity-Presentationidentity credentials presentpresentation → Identity-Presentation

An AAT proves issuance by Link without disclosing user claims. A presentation discloses selected claims and proves possession of the credential's holder key for an audience and nonce. If the service requires both, send both headers.

Setup and handling

Use Link Agent Wallet v0.27.0 or later. Check link-cli --version; if older or missing, install or update with npm install -g @stripe/link-cli@latest. No LINK_IDENTITY_COMMANDS flag is needed. See wallet setup; reuse the create-payment-credential skill's authentication guidance.

  • Identity commands are available through the CLI, not its MCP tools.
  • Only request needs Link authentication. list, pop, and present use local files. Requests return paths and metadata; pop and present return proofs.
  • Capture stdout with --format json, parse it programmatically with a JSON parser or jq, and pass the exact proof to the HTTP client or browser automation. Never manually transcribe, reconstruct, or re-encode tokens or presentations. Keep proofs, private keys, and claim values out of transcripts and logs; use restrictive permissions for handoff files.
  • Cached identity files belong to the home directory. Changing --auth or logging in as another user does not switch the cached identity. Use separate home directories for different users.

Attestations

# Inspect inventory.
link-cli identity attestations list --format json

# Refill only when needed. Count is required: 1–100.
link-cli identity attestations request --count 10 --format json

# Pop a token when ready to use it.
link-cli identity attestations pop --format json

Check errors and attestations. Only entries with storage: "pool" are available to pop; total_token_count also includes exports. Requests append to ~/.link-cli/identity/attestations/pool.json.

Example pop output:

{
  "issuer": "https://api.link.com",
  "token_key_id": "<issuer-key-id>",
  "token": "<token>",
  "authorization": "PrivateToken token=\"<token-with-header-padding>\""
}

Use authorization verbatim, including its scheme and quoting. pop removes the token locally; server reuse rules belong to the service. Keep the same AAT for retries tied to an existing interaction, and use a fresh one for a new service. Do not return popped tokens to the pool.

For agent-managed allocation, export a batch instead:

link-cli identity attestations request --count 10 --output-file ./aats.json --format json

Use a new file outside ~/.link-cli/identity/attestations. Read its tokens array and each entry's authorization programmatically. Exported tokens never enter the CLI pool; the caller owns allocation, concurrency, and cleanup.

Credentials and presentations

# Check errors, expired, expires_at, and claim_names.
link-cli identity credentials list --format json

# Request only if no usable credential exists for the intended user.
link-cli identity credentials request --format json

Issuance replaces ~/.link-cli/identity/credentials/current.json and creates or reuses ~/.link-cli/identity/holder-key.jwk. Keep both local. Choose claims with present:

link-cli identity credentials present \
  --aud 'https://service.example' \
  --nonce '<nonce-from-challenge>' \
  --claim email \
  --format json

Use the service challenge's exact audience and nonce after checking that they belong to the intended service. All three options are required. Repeat --claim only for additional requested claims authorized by the user's task. Pass challenge values as individual process arguments or quote them safely in a shell.

Example output:

{"presentation":"<issuer-jwt>~<email-disclosure>~<key-binding-jwt>"}

present signs locally with the saved holder key and leaves the files unchanged. Send the presentation promptly. Reuse the underlying credential until expiry, but obtain a fresh challenge after a nonce is consumed or expires.

Send and recover

Set headers only for the intended endpoint. Preserve the challenged operation's method, body, session, interaction identifiers, and any required AAT. Avoid global browser headers or redirects that forward proofs to another origin. Report success only after the service accepts the request.

  • ATTESTATION_POOL_EMPTY: request a batch, then pop a token.
  • Errors in list: inspect the reported file/error instead of treating storage as empty.
  • Unavailable claim: check claim_names; do not disclose extra claims or the raw credential.
  • Missing or mismatched holder key: preserve the files and restore the correct key or explicitly obtain a new credential.
  • Audience or nonce rejection: check the intended service and its current challenge before presenting again.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.