← Back to Skills Marketplace
daav3

Agentic Lending EVK

by dave · GitHub ↗ · v0.1.1 · MIT-0
cross-platform ⚠ suspicious
69
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install agentic-lending-evk
Description
Plan and execute the current EVK-first agentic lending workflow for Api3-backed markets. Use when a user wants an agent to discover or fund the required Api3...
README (SKILL.md)

Agentic Lending EVK

This skill turns the repo's current planner primitives into one agent-facing workflow, from feed readiness through deployment and live borrow proof.

Safety expectations

  • default to planning, simulation, or preview mode first
  • treat browser-assisted funding as optional and only automate it when the user wants the flow completed end to end
  • only use live signer-backed funding, deployment, or borrow canaries when the user explicitly asks for live execution
  • for the bundled borrow-proof script, require both --live and config.liveAcknowledgement === "I_UNDERSTAND_THIS_WILL_SEND_TRANSACTIONS"
  • require the configured proof account address to match the signer address before any live proof transaction is sent
  • keep private key material in environment variables or local runtime config, never in committed files

Use this when

  • the user wants to stand up or dry-run an EVK lending market backed by Api3 feeds
  • the user wants one workflow instead of manually stitching planner JSON steps
  • the task includes any combination of asset selection, feed readiness, feed funding, oracle preparation, EVK deployment prep, guarded dry-run or send, or post-deploy borrow proof

Do not use this when

  • the user wants Morpho, not EVK
  • the requested recipe or oracle path is not supported by the current repo
  • the task requires pretending a skeleton-only path is executable
  • the task assumes fully autonomous feed funding in every case

Read next

  1. references/api_reference.md
  2. references/current_capabilities.md
  3. references/live-borrow-checklist.md when the workflow reaches post-deploy proof
  4. references/arbitrum-eusdc1-isolated-example.json when a concrete proof config shape helps
  5. Run node scripts/evk_live_borrow_proof.js --help for the bundled proof executor

Minimum required inputs for the combined flow

Gather these before you assume the full combined workflow is feasible:

  • chain name and chain id
  • collateral asset set and borrow asset set
  • target recipe id or risk preset
  • whether the user wants planning only, dry-run, or real send
  • whether the user also wants post-deploy borrow proof
  • live RPC availability for the chain
  • signer expectations for any live funding, deployment, or proof step
  • vaultContext when the deployment path needs concrete factory or asset wiring

For a full live end-to-end run, also confirm:

  • feed funding branch expectations, especially whether fundingExecutionClassification.state may be browser-assisted
  • browser access if the Api3 Market flow must be automated
  • a safe tiny borrow amount for the proof step
  • a target collateral vault for the proof, not just the borrow vault

If any of those are missing, stop and ask instead of improvising.

Default workflow

  1. Normalize the request into:
    • chain
    • collateral assets
    • borrow assets
    • target recipe or risk preset
    • whether the user wants planning only, dry-run, or real send
    • whether the user also wants proof of actual borrowability
  2. Prefer run-evk-workflow as the default end-to-end entrypoint when the user wants the full EVK path.
    • pass recipeId explicitly when known, otherwise let the runner infer from the current EVK recipe catalog
    • treat status as the canonical summary object
    • treat result as the detailed staged artifact bundle
  3. If deployment succeeds and the user wants proof of actual borrowability, continue within this skill to the built-in post-deploy proof step.
    • treat deployment success and borrowability proof as separate milestones
    • use the deployment outputs plus live vault addresses, collateral-vault targets, EVC state, and tiny borrow sizing to build the proof config
    • require preview mode first, then live canary only if the user wants a real proof
  4. Fall back to the manual staged commands only when the user asked for a subset, wants intervention at a specific phase, or the runner stops before the desired outcome.
    • plan-market
    • ensure-feeds
    • if feeds are live, continue to the oracle and EVK planner path
    • if funding is needed, inspect purchase-inputs or prepare-buy-subscription output and branch on fundingExecutionClassification.state
    • if fundingExecutionClassification.state === "executable":
      • purchase-inputs
      • prepare-contract-call or prepare-buy-subscription
      • execute-buy-subscription
      • ensure-feeds or ensure-active again to confirm the feed is now live
    • if fundingExecutionClassification.state === "browser-assisted":
      • call browser-plan
      • use the browser tool to execute the returned plan if the required UI is reachable and the user wants full automation
      • re-run ensure-feeds or ensure-active after the browser flow
    • if fundingExecutionClassification.state === "unsupported", stop and report unsupported status clearly
    • prepare-euler-oracle
    • prepare-evk-market
    • prepare-evk-deployment
    • deploy-evk-market
    • scripts/evk_live_borrow_proof.js for the post-deploy proof step when requested
  5. Stop and report honestly if:
    • feed funding is unsupported
    • the browser-assisted funding UI is not reachable or cannot be automated safely
    • the oracle path is still skeleton-only
    • the EVK market path is still skeleton-only
  6. Prefer dry-run first for exact onchain funding and deployment paths.
  7. For browser-assisted funding, prefer browser automation over asking the user to click through manually when the UI is reachable.
  8. Only use real send when the user explicitly wants it and the guarded acknowledgement contract is present.
  9. Do not claim a market is borrowable just because deployment succeeded. A live or at least previewed post-deploy proof step is required for that claim.

Built-in post-deploy borrow proof

Use scripts/evk_live_borrow_proof.js.

It supports:

  • preview mode with a read-only account address
  • live mode with a signer loaded from an environment variable
  • optional debt repayment before the new proof
  • optional single-hop Uniswap V3 style swaps to source repay or collateral assets
  • collateral deposit into the target EVK vault
  • disabling stale collateral vaults in EVC
  • disabling stale controllers in EVC
  • enabling the target collateral and optionally the target controller
  • final tiny borrow and before or after logging

Keep the proof honest:

  • distinguish mixed-collateral borrow from isolated target-collateral borrow
  • preserve the exact funding branch that led here
  • verify quoteability, LTV wiring, EVC state, and vault cash before a live canary
  • if the chain does not use a Uniswap V3 compatible router, replace or bypass the bundled swap step explicitly

Current happy-path rule

The currently supported end-to-end EVK path requires:

  • feeds that are already live, or a feed that can be activated through the current narrow exact buySubscription(...) path
  • an oracle path that already resolves to executable oracle transactions
  • top-level prepare-evk-deployment.vaultContext
    • factoryAddress
    • assetAddress
  • guarded send fields on the funding and deployment executor requests

Without that vaultContext, the first EVK market transaction remains skeleton-only and readyToBroadcast must stay false. Without a supported exact buy-subscription path, feed funding must stay at handoff or prep only.

Output expectations

When using this skill, always report:

  • status.phaseReached when the top-level runner is used
  • status.state when the top-level runner is used
  • status.recipeId when the top-level runner is used
  • what is executable vs still skeleton-only
  • blockers and warnings
  • fundingExecutionClassification.state when feed funding is involved manually
  • status.fundingExecutionState when the top-level runner is used
  • whether the current state is:
    • plan-only
    • feed-ready but awaiting activation or funding handoff
    • funding dry-run ready
    • browser-assisted funding ready
    • oracle-executable
    • EVK dry-run ready
    • real-send ready
    • deployed but borrow-proof pending
    • live borrow proven

If the user asked for borrowability proof, also report whether the workflow has already reached the post-deploy proof step or is still waiting on proof inputs.

If the top-level runner is used, also surface the executable flags under status.executable instead of paraphrasing them away.

Tone

Be blunt about current limits. Do not overclaim generalized support. Do not blur deployment success into borrowability proof. If the workflow falls off the happy path, say exactly where and why.

Usage Guidance
This skill looks like what it claims: a workflow + a bundled Node script that can dry-run or send real Ethereum (Arbitrum) transactions. Important things to consider before installing or running it: - Protect your private key: only set ARBITRUM_PRIVATE_KEY (or whatever privateKeyEnv you use) in the environment if you intend to perform live sends. The skill's docs say preview/dry-run is supported without a key — avoid putting production keys in persistent shells/CI unless necessary. - Review and keep configs off Git: borrow-proof-config.json and any request.* JSONs may reference real accounts, RPC URLs, and privateKeyEnv names. Keep any secret-bearing config out of source control. - Use preview/dry-run first: the skill explicitly prefers planning and dry-run. Run the script with --config <path> (no --live) to see the exact transactions it would submit. - Confirm the exact broadcast rules: SKILL.md requires explicit broadcast flags and an acknowledgement token before real sends; ensure the agent actually enforces those before allowing a live run. - Browser automation: the workflow can automate browser-assisted funding flows if you allow it — that grants automation broad control of web UIs; be cautious about enabling that on behalf of your account. - Consider using a limited ephemeral account for live proofs: fund a throwaway/test wallet with just enough to run the tiny borrow canary, rather than exposing a high-value key. The main coherence issue is that the registry metadata forces ARBITRUM_PRIVATE_KEY as a required env var even though the documented default behavior supports preview without a key; this is more restrictive than necessary but not itself malicious. If you need stronger assurance, inspect the full evk_live_borrow_proof.js file and the planner binaries referenced by the skill before enabling live runs.
Capability Analysis
Type: OpenClaw Skill Name: agentic-lending-evk Version: 0.1.1 The skill bundle facilitates automated DeFi lending operations on Arbitrum, including market deployment and 'canary' borrowing. It requires access to a private key (ARBITRUM_PRIVATE_KEY) and includes a Node.js script (scripts/evk_live_borrow_proof.js) that performs high-risk actions such as infinite token approvals (MaxUint256), Uniswap V3 swaps, and borrowing. The instructions in SKILL.md and api_reference.md explicitly direct the agent to use a 'browser tool' for 'browser-assisted funding' of Api3 feeds, which represents a significant privilege level and attack surface. While the documentation emphasizes dry-runs and user consent, the combination of financial transaction capabilities, private key handling, and automated browser interaction constitutes a high-risk profile.
Capability Tags
cryptorequires-walletcan-make-purchasesrequires-sensitive-credentials
Capability Assessment
Purpose & Capability
Name, description, included docs, and the evk_live_borrow_proof.js script all align: this repo-level skill is genuinely for planning, funding (including browser-assisted flows), deploying EVK markets, and running a post-deploy borrow canary. The declared required config files match the CLI usage examples. One mismatch: the registry metadata marks ARBITRUM_PRIVATE_KEY as a required env var (primaryEnv), but the SKILL.md explicitly recommends defaulting to planning/preview and states live signer-backed execution should only be used when the user explicitly asks for it. The private key is necessary for live sends but not for dry-run/preview, so listing it as a hard requirement is stricter than the documented operational model.
Instruction Scope
The SKILL.md instructions are detailed and scoped to the EVK-first workflow. They instruct the agent to prefer plan/dry-run modes, only use live sends when requested and when guarded acknowledgement and signer info are present, and to use browser automation only for browser-assisted funding when the user wants full automation. The instructions reference only the declared config files and the bundled script; there are no directives to read unrelated system files or exfiltrate data.
Install Mechanism
This is an instruction-only skill with included JS files and no install spec. That keeps risk lower than arbitrary downloads. The code executes under the user's node runtime; there are no external install URLs or package downloads declared in the skill metadata.
Credentials
The skill declares a single primary credential ARBITRUM_PRIVATE_KEY which is appropriate for live transaction signing. However, the SKILL.md and scripts support preview mode without a private key (they read privateKeyEnv from the config and only require it for --live). Marking ARBITRUM_PRIVATE_KEY as a required env var for the skill's installation or invocation is inconsistent with the documented preview-first workflow and is more permissive than necessary for planning/dry-run operations. Users should not set long-lived production keys in the environment unless they intend to perform live transactions.
Persistence & Privilege
always:false and normal agent-invocation settings are used. The skill does not request persistent/always-on inclusion and does not modify other skills or system settings. The primary privilege is the ability to sign/send on-chain transactions when a private key is supplied — this is expected for its purpose but is a high-impact capability that requires user caution.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install agentic-lending-evk
  3. After installation, invoke the skill by name or use /agentic-lending-evk
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.1
- Added support and safety guidance for browser-assisted Api3 Market feed funding. - Updated skill metadata with author, homepage, environment, and config requirements. - Clarified default behavior: prefer planning, simulation, or preview mode before any live onchain execution. - Strengthened safety expectations: only perform live funding or deployment steps when explicitly requested by the user. - Extended documentation to reflect browser-based funding automation as an option when available and user-approved.
v0.1.0
- Initial release of agentic-lending-evk skill. - Provides a complete agent-facing workflow for EVK lending market deployment with Api3-backed oracles, including feed funding, oracle preparation, market assembly, deployment, and post-deploy live borrow proof. - Explicitly limited to the current EVK-first, Api3 feed-backed, and supported recipe paths—excludes Morpho and unsupported oracle recipes. - Workflow includes clear status reporting, executable state tracking, and step-by-step handling of feed funding, browser-assisted flows, and fallback behavior. - Built-in support for honest, verifiable post-deploy borrow proofs using the included script, with robust reporting of blockers and workflow limits.
Metadata
Slug agentic-lending-evk
Version 0.1.1
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is Agentic Lending EVK?

Plan and execute the current EVK-first agentic lending workflow for Api3-backed markets. Use when a user wants an agent to discover or fund the required Api3... It is an AI Agent Skill for Claude Code / OpenClaw, with 69 downloads so far.

How do I install Agentic Lending EVK?

Run "/install agentic-lending-evk" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Agentic Lending EVK free?

Yes, Agentic Lending EVK is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Agentic Lending EVK support?

Agentic Lending EVK is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Agentic Lending EVK?

It is built and maintained by dave (@daav3); the current version is v0.1.1.

💬 Comments