About

Imported from starchild-ai-agent/official-skills (1inch/SKILL.md). Install upstream with npx skills add starchild-ai-agent/official-skills --skill 1inch. Copyright stays with the author.

1inch (Script-First, Install-and-Use)

Name: 1inch
Author: starchild-ai-agent

This skill is designed for your architecture: do not rely on platform-injected tools.

✅ Uses only skill-local scripts under skills/1inch/scripts/
✅ Agent executes scripts via bash (python3 ...)
✅ No oneinch_* tool calls required

Why this version fixes “tool not found”

Old design depended on runtime tool registration (oneinch_quote, oneinch_swap, ...). If tool injection fails, agent is blocked.

New design uses deterministic local scripts:

call 1inch HTTP API directly
call wallet service directly (OIDC via /.fly/api)
print JSON result

So after install, the agent always has a runnable path.

Files

scripts/_oneinch_lib.py — shared client + wallet/OIDC helpers
scripts/tokens.py — token search/list
scripts/quote.py — quote only
scripts/check_allowance.py — allowance check
scripts/approve.py — approve tx broadcast
scripts/swap.py — swap execution (optional auto-approve)
scripts/run_swap_flow.py — one-command flow (quote + optional approve + swap + post-trade balance verification)

Environment

Required:

ONEINCH_API_KEY
WALLET_SERVICE_URL (default fallback exists)
sc-proxy connectivity for 1inch API (mandatory)

Assumptions:

Running on Fly Machine with /.fly/api unix socket for OIDC token minting.

sc-proxy requirement (critical)

This skill enforces 1inch calls through sc-proxy.

It reads proxy from HTTP_PROXY/HTTPS_PROXY first, else falls back to PROXY_HOST + PROXY_PORT.
If no proxy env is found, scripts fail fast with a clear error instead of silently direct-connecting.

Supported chains

ethereum, arbitrum, base, optimism, polygon, bsc, avalanche, gnosis

Agent execution rules (IMPORTANT)

When user asks for 1inch actions, follow this exact pattern:

Never call oneinch_* tools
Run local scripts with bash + python3 skills/1inch/scripts/<script>.py ...
Parse JSON output and respond with concise summary
For write actions, always do post-check using wallet balance tools (if available) or rerun script-based checks

Canonical mapping

“查 token 地址 / 列 token” → tokens.py
“先报价 / 看能换多少” → quote.py
“检查授权” → check_allowance.py
“授权” → approve.py
“执行兑换” → swap.py

Command examples

1) Search token

python3 skills/1inch/scripts/tokens.py --chain polygon --search POL --limit 10

2) Quote: 1 USDC -> POL

python3 skills/1inch/scripts/quote.py --chain polygon --from USDC --to POL --amount 1

3) Check allowance (USDC)

python3 skills/1inch/scripts/check_allowance.py --chain polygon --token USDC

4) Approve USDC (unlimited)

python3 skills/1inch/scripts/approve.py --chain polygon --token USDC

5) Swap: 1 USDC -> POL (auto-approve if needed)

python3 skills/1inch/scripts/swap.py --chain polygon --from USDC --to POL --amount 1 --slippage 1.0 --auto-approve

6) One-command full flow (recommended)

python3 skills/1inch/scripts/run_swap_flow.py --chain polygon --from USDC --to POL --amount 1 --slippage 1.0 --auto-approve

Tune retries when needed:

python3 skills/1inch/scripts/run_swap_flow.py --chain polygon --from USDC --to POL --amount 1 --slippage 1.0 --auto-approve --swap-retries 3 --swap-retry-backoff 2

Default workflow for “买 X USDC 的 POL”

Preferred deterministic sequence:

run_swap_flow.py --auto-approve
Read JSON result and report quote, tx submission result, and verification deltas

Fallback manual sequence (if user asks step-by-step):

quote.py (confirm expected output)
swap.py --auto-approve
Verify with fresh balances (before/after)

If swap returns wallet policy rejection:

load wallet-policy skill
propose wildcard baseline (DENY exportPrivateKey, ALLOW *)
after user confirms, rerun swap command

Error handling

Unknown chain → ask user to choose supported chain
1inch API 4xx/5xx → show raw error; run_swap_flow.py automatically retries transient /swap 5xx (default: 2 retries with exponential backoff)
Not enough <token> balance from 1inch → likely symbol mapped to a different token variant (e.g. USDC.e vs USDC); rerun with explicit token address from tokens.py
Wallet API 4xx/5xx → show raw error, do not fabricate tx hash
insufficient allowance (without auto-approve) → rerun with --auto-approve or run approve.py
policy rejection → propose policy update then retry

Notes

Amount input is human units (e.g. --amount 1 = 1 USDC), script handles decimal conversion.
Token symbol resolution comes from 1inch /tokens on the selected chain.
If a symbol has multiple variants on a chain (e.g., USDC / USDC.e), prefer passing token contract address explicitly to avoid ambiguity.
For native token input, use native / ETH.

Quick smoke test

python3 skills/1inch/scripts/quote.py --chain polygon --from USDC --to POL --amount 1

If this works, the skill is operational in script mode.

Script Reference (CLI args)

This skill exposes scripts (not Python functions). Run them via bash + python3 skills/1inch/scripts/<name>.py <args>. All scripts emit JSON on stdout — parse with json.loads(...) or just print the output.

Common arguments

Arg	Required	Notes
`--chain`	yes	One of `ethereum`, `arbitrum`, `base`, `optimism`, `polygon`, `bsc`, `avalanche`, `gnosis`
`--from`, `--to`	yes (swap/quote)	Token symbol (`USDC`, `POL`, etc.) or contract address. Use `native` / `ETH` for native gas token.
`--amount`	yes (swap/quote)	Human units (e.g. `1` = 1 USDC, decimal conversion handled internally)
`--slippage`	optional	Percent (default `1.0`)
`--token`	yes (allowance/approve)	Token symbol or address
`--auto-approve`	optional flag	swap/run_swap_flow only — auto-broadcasts approve if allowance insufficient
`--swap-retries`	optional	run_swap_flow only — retries on `/swap` 5xx (default `2`)
`--swap-retry-backoff`	optional	run_swap_flow only — backoff seconds (default `1`)

Scripts

Script	Purpose	Required args
`tokens.py`	List or search tokens on a chain	`--chain` + `--search SYM` and/or `--limit N`
`quote.py`	Get quote (no broadcast)	`--chain --from --to --amount`
`check_allowance.py`	Check ERC-20 allowance to 1inch router	`--chain --token`
`approve.py`	Broadcast approve tx (unlimited)	`--chain --token`
`swap.py`	Execute swap	`--chain --from --to --amount` (+ optional `--slippage`, `--auto-approve`)
`run_swap_flow.py`	One-command quote + (auto-approve) + swap + post-trade balance check	Same as `swap.py` (+ optional retry args)

Recommended flow

For ANY user "swap X for Y" request, default to run_swap_flow.py --auto-approve. It handles the full chain end-to-end and verifies via wallet balance read after broadcast.

If user asks "just quote, don't execute" → quote.py only.

Output schema (run_swap_flow.py)

JSON with these top-level keys:

quote — {fromToken, toToken, fromAmount, toAmount, estimatedGas, ...}
approve — {required, txHash, success} (only if --auto-approve)
swap — {txHash, success, error?}
verification — {from_balance_before, from_balance_after, to_balance_before, to_balance_after, delta_in, delta_out}

If wallet returns a policy rejection, swap.success = false and swap.error will mention policy. Load the wallet-policy skill, propose a wildcard baseline, then rerun.

Use it

Copy one of these into your project. Installing also returns the manifest and these snippets.

yaml

targets:
  - https://api.opensmartroute.ai/api/v1/registry/starchild-ai-agent-official-skills-1inch/manifest   # or paste the manifest below

Manifest

An Open Capability Manifest: the router reads it to know what this does, what it costs and when to pick it.

starchild-ai-agent-official-skills-1inch.ocm.jsonjson

{
  "ocm": "1",
  "id": "starchild-ai-agent-official-skills-1inch",
  "kind": "skill",
  "name": "1inch",
  "description": "1inch DEX aggregator: same-chain EVM swaps, SOL↔EVM cross-chain, limit orders. Use when swapping tokens with best routing or placing limit orders (e.g. ETH→USDC on Base, SOL→ETH cross-chain, limit buy ARB at 0.80).",
  "publisher": "starchild-ai-agent",
  "version": "1.0.0",
  "capabilities": {
    "domains": [
      "math"
    ],
    "tags": [
      "skill-md",
      "1inch",
      "dex",
      "swap",
      "evm",
      "script-first",
      "skills-sh"
    ],
    "languages": [
      "en"
    ]
  },
  "quality_prior": 0.6,
  "examples": [
    "1inch DEX aggregator: same-chain EVM swaps, SOL↔EVM cross-chain, limit orders. Use when swapping tokens with best routing or placing limit orders (e.g. ETH→USDC on Base, SOL→ETH cross-chain, limit buy ARB at 0.80)."
  ],
  "primary": false,
  "metadata": {
    "source": {
      "provider": "skills.sh",
      "repository": "https://github.com/starchild-ai-agent/official-skills",
      "path": "1inch/SKILL.md",
      "ref": "HEAD",
      "url": "https://github.com/starchild-ai-agent/official-skills/blob/HEAD/1inch/SKILL.md",
      "key": "starchild-ai-agent/official-skills/1inch/SKILL.md"
    }
  },
  "instructions": "# 1inch (Script-First, Install-and-Use)\n\nThis skill is designed for your architecture: **do not rely on platform-injected tools**.\n\n- ✅ Uses only skill-local scripts under `skills/1inch/scripts/`\n- ✅ Agent executes scripts via `bash` (`python3 ...`)\n- ✅ No `oneinch_*` tool calls required\n\n---\n\n## Why this version fixes “tool not found”\n\nOld design depended on runtime tool registration (`oneinch_quote`, `oneinch_swap`, ...).\nIf tool injection fails, agent is blocked.\n\nNew design uses deterministic local scripts:\n\n1. call 1inch HTTP API directly\n2. call wallet service directly (OIDC via `/.fly/a",
  "cost": {
    "context_tokens": 1916
  }
}

Fetch it by URL: GET /api/v1/registry/starchild-ai-agent-official-skills-1inch/manifest?version=1.0.0

Reviews

Star ratings from people who tried it. One review per account; edit yours any time.

No reviews yet. Install it, try it, and be the first to rate it.