agentsclimarketplace

Pay recover

Skill runxhq/runx/skills/spend/graph/pay-recover

Reconcile an idempotent payment attempt before retrying or sealing.From its SKILL.md

Install
npx -y skills add runxhq/runx --skill pay-recover

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

SKILL.md

1.5 KB, 322 tokens by cl100k_base, as published. Nobody here has run it

Pay Recover

Inspect a payment idempotency key after a crash, timeout, retry, or ambiguous rail response.

This skill is the recovery surface for agent payments. It answers one question: has this reserved payment already reached a rail outcome that can be sealed, or is a retry still safe? It must prefer reconciliation over mutation.

It does not spend. It does not decide success without a proof ref. It reports ambiguous states as escalation.

Tie every conclusion to the reservation decision, idempotency key, rail profile, and proof references. The safe outcomes are deliberately narrow: seal recovered proof, retry once under the same key, decline, or escalate. If the rail cannot prove success, failure, or a safe same-key retry, return escalate.

Output

  • recovery_assessment: recovered, retry_safe, failed, or ambiguous.
  • rail_lookup: what was queried and which idempotency key was used.
  • proof_refs: recovered rail proof refs, if any.
  • recommended_action: seal, retry_same_key, decline, or escalate.
  • open_questions: unresolved state that blocks safe execution.

Inputs

  • idempotency (required): reservation key and recovery lookup fields.
  • reserved_payment_authority (required): child payment authority term.
  • rail_profile_ref (required): configured rail profile reference.
  • prior_rail_result (optional): previous rail attempt result.
  • receipt_refs (optional): existing harness or rail proof refs.

What ships with it: 1 file

4.2 KB alongside SKILL.md

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.