agentsclimarketplace

Pay fulfill rail

Skill runxhq/runx/skills/spend/graph/pay-fulfill-rail

Fulfill a reserved payment challenge through one rail under attenuated runx authority.From its SKILL.md

Install
npx -y skills add runxhq/runx --skill pay-fulfill-rail

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

3.3 KB, 680 tokens by cl100k_base, as published. Nobody here has run it

Pay Fulfill Rail

Execute one rail operation below the runx spend gate.

This skill adapts a protocol or provider challenge to the credential/proof shape needed by the paid tool. It can spend only when the parent harness has already selected a Decision, reserved budget by idempotency key, and passed an attenuated payment authority term into the child harness.

The skill must receive a scoped spend capability or provider session reference, never raw funding material. It returns rail proof for the receipt; it does not decide policy, approval, retry, or success.

Bind the provider response, challenge id, idempotency key, amount, currency, counterparty, and proof hash or reference into the rail evidence while redacting sensitive fields. Report only operational states—fulfilled, declined, retryable, recovered, or ambiguous. Return needs_agent or ambiguous when the response cannot be tied to both the reserved authority and idempotency key.

There is intentionally no x402 fulfillment runner here. The public x402-pay facade is plan-only until a trusted buyer adapter implements the standard paid resource retry and returns independent settlement readback. Do not substitute agent-authored evidence or caller-supplied signer/facilitator endpoints.

Output

  • rail_result: rail status, amount, currency, counterparty, and operation.
  • rail_proof: redacted proof payload or proof ref for the child harness receipt.
  • credential_envelope: credential or token returned to the paid tool, with sensitive fields redacted or referenced.
  • redactions: fields withheld from receipts and logs.
  • recovery_hint: idempotency/retry guidance for pay-recover.

Inputs

  • payment_challenge (required): protocol/provider challenge to fulfill.
  • reserved_payment_authority (required): child payment authority term.
  • spend_capability_ref (required): scoped single-use spend capability ref.
  • rail_profile_ref (required): configured rail profile reference.
  • payment_admission (optional): hosted payment admission token and settlement identity. When present, it is bound into supervisor settlement evidence.
  • idempotency (required): reservation key and recovery fields.
  • quote_packet (optional): source quote packet for evidence continuity.

Agent task contracts

pay-fulfill-rail-mock

Produce deterministic mock rail evidence only for an explicitly configured test profile. Bind the challenge, reserved authority, capability reference, rail profile, and idempotency key. Return rail_result, rail_proof, a redacted credential_envelope, redactions, and recovery_hint. Never claim a live provider call or real money movement, and never accept raw funding material.

pay-fulfill-rail-mpp

Interpret only bounded MPP provider evidence supplied through the authorized rail context. The same five outputs must bind amount, currency, counterparty, operation, admission/capability, and idempotency. Report fulfilled only when terminal provider evidence matches the reservation; otherwise return declined, retryable, recovered, ambiguous, or needs_agent. Never expose credentials or turn acknowledgement alone into settlement finality.

What ships with it: 3 files

17.1 KB alongside SKILL.md, 1 of them executable

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.