
wallet-policy
Generate Privy wallet policy rules from natural language. Use when the user wants to set up, modify, or review wallet security policies — transfer limits, address allowlists, method restrictions, time windows, etc.
- Installs
- 4.1K
- Last 24h
- —
- Trending rank
- —
- Security
- Warn
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/starchild-ai-agent/official-skills --skill wallet-policy- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads wallet-policy automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
Security audits
Independent scans of the skill's instructions and bundled files.
SKILL.md
The instructions your agent reads when the skill is triggered.
Wallet Policy Generator
You help users create wallet security policy rules. The user describes what they want in plain language, and you generate the exact Privy policy rules JSON. After generating the rules, you MUST call the wallet_propose_policy tool to send the proposal to the user for review and approval.
Always respond in the user's language.
Output Format
After generating the policy rules, call the wallet_propose_policy tool:
wallet_propose_policy(
chain_type="ethereum", # "ethereum" or "solana"
title="Update EVM Wallet Policy",
description="Allow transfers to treasury address",
rules=[
{
"name": "Allow transfers to treasury",
"method": "eth_sendTransaction",
"conditions": [
{
"field_source": "ethereum_transaction",
"field": "to",
"operator": "eq",
"value": "0x1234567890abcdef1234567890abcdef12345678"
}
],
"action": "ALLOW"
}
]
)
The tool sends an action_request event to the frontend, which displays the proposed policy to the user for confirmation. The user must approve (and sign) before the policy is applied. Do NOT output rules as code blocks — always use the tool.
If the user's request covers both EVM and Solana, call wallet_propose_policy twice — once with chain_type="ethereum" and once with chain_type="solana".
CRITICAL — Tool invocation is mandatory:
- You MUST call
wallet_propose_policyfor EVERY policy request. Never output rules as plain text or code blocks. - For dual-chain requests (both EVM and Solana), call the tool TWICE — once per chain_type.
- The tool validates rules against the Privy API schema. If validation fails, fix the errors and retry.
Policy Engine Basics
Tell the user these fundamentals when relevant:
- Default allow-all — New wallets have NO policy (all transactions allowed). Policy is opt-in. Once a policy is attached, it switches to deny-by-default: any request that matches no rules is denied. An empty rules array = deny everything.
- DENY wins — If any DENY rule matches, the request is blocked even if ALLOW rules also match.
- Multiple conditions = AND — All conditions in a single rule must match for the rule to trigger.
- Multiple rules = evaluated in order — First matching DENY blocks; otherwise first matching ALLOW permits.
- Solana per-instruction — Every instruction in a Solana transaction must individually match an ALLOW rule.
Constructing Policy Rules
Default Approach: Wildcard Policy
For any on-chain service (Hyperliquid, Orderly, 1inch, or any new dapp), propose the standard wildcard policy:
wallet_propose_policy(
chain_type="ethereum",
title="Enable Wallet Operations",
description="Allows all transactions and signing on all EVM chains. Only blocks private key export. The user signs each individual transaction for approval.",
rules=[
{
"name": "Deny key export",
"method": "exportPrivateKey",
"conditions": [],
"action": "DENY"
},
{
"name": "Allow all operations",
"method": "*",
"conditions": [],
"action": "ALLOW"
}
]
)
This works because:
- The wallet policy acts as a capability gate — the user's signature on the policy is explicit consent to enable on-chain operations
- Individual transactions still require user approval in the frontend before execution
- The DENY on
exportPrivateKeyprevents the most dangerous operation (key extraction) - The
*wildcard covers all transaction types, signing methods, and chains — no service-specific rules needed
When to use specific rules instead: Only when the user explicitly requests tighter restrictions (e.g. "only allow transfers under 1 ETH", "only allow transactions on Arbitrum", "only allow this specific contract address"). In that case, use the rule-building reference below.
Building Custom Restrictive Rules
If the user wants tighter control, identify what transactions the service needs:
- What contract addresses will be called? (the
tofield) - What chain will it operate on? (the
chain_id) - What value will be sent? (native token amount in wei)
- Does it need EIP-712 signing? (typed data for off-chain orders, permits)
- Does it need token approvals? (ERC-20 approve calls to token contracts)
Map each transaction type to a policy rule:
| Transaction type | Rule pattern |
|---|---|
| Call a specific contract | ethereum_transaction.to = contract address + chain_id = chain |
| ERC-20 token approval | ethereum_transaction.value = "0" + chain_id = chain (approvals are zero-value calls to the token contract) |
| EIP-712 typed data signing | ethereum_typed_data_domain.verifyingContract = contract address |
| Any transaction on a chain | ethereum_transaction.chain_id = chain |
| Smart contract deployment | Use wildcard pattern (deployments have no fixed to address) |
Propose and Explain
Always use wallet_propose_policy to send the proposal to the user. In the description field, explain:
- What the rules allow
- What security tradeoffs exist (e.g. wildcard allows all operations, but each tx still requires user approval)
Complete Rule Schema
{
"name": "string (1-50 chars, descriptive)",
"method": "<method>",
"conditions": [ <condition>, ... ],
"action": "ALLOW" | "DENY"
}
Supported Methods
| Method | Chain | Description |
|---|---|---|
eth_sendTransaction | EVM | Broadcast a transaction |
eth_signTransaction | EVM | Sign without broadcasting |
eth_signTypedData_v4 | EVM | Sign EIP-712 typed data |
eth_signUserOperation | EVM | Sign ERC-4337 UserOperation |
eth_sign7702Authorization | EVM | EIP-7702 authorization |
signTransaction | Solana | Sign a Solana transaction |
signAndSendTransaction | Solana | Sign and broadcast |
signTransactionBytes | Tron/SUI | Sign raw transaction bytes |
exportPrivateKey | Any | Export the private key |
* | Any | Wildcard — matches all methods |
Note: personal_sign (message signing) and signMessage (Solana) are NOT valid policy methods. They cannot be individually allowed/denied. To allow message signing, use * wildcard. Under deny-all (empty rules), message signing is also blocked.
Condition Object
{
"field_source": "<source>",
"field": "<field_name>",
"operator": "<op>",
"value": "<string>" | ["<string>", ...]
}
Operators:
eq— equals (single value)gt,gte,lt,lte— comparison operators (numeric string values)in— matches any value in array (max 100 values). Use this for multiple addresses/values.
Do NOT use in_condition_set:
in_condition_set— This operator requires pre-created condition sets via Privy API, which you cannot create. Always use theinoperator instead for arrays of addresses or values. If you need more than 100 values, split into multiple rules.
Examples:
// ✅ CORRECT: Multiple addresses with "in" operator
{"field": "to", "operator": "in", "value": ["0xAddr1...", "0xAddr2...", "0xAddr3..."]}
// ❌ WRONG: Do NOT use "in_condition_set" - you cannot create condition sets
{"field": "to", "operator": "in_condition_set", "value": "a2p4etpcbj2dltbjfigybi8j"}
{"field": "to", "operator": "in_condition_set", "value": ["0xAddr1...", "0xAddr2..."]}
// ✅ CORRECT: For many addresses, use multiple rules with "in" operator
// Rule 1: First 100 addresses
{"field": "to", "operator": "in", "value": ["0xAddr1...", "0xAddr2...", /* ... 100 addresses */]}
// Rule 2: Next batch
{"field": "to", "operator": "in", "value": ["0xAddr101...", "0xAddr102...", /* ... */]}
Condition Types Reference
1. ethereum_transaction
Fields: to, value, chain_id
{"field_source": "ethereum_transaction", "field": "to", "operator": "eq", "value": "0xAbC..."}
{"field_source": "ethereum_transaction", "field": "value", "operator": "lte", "value": "1000000000000000000"}
{"field_source": "ethereum_transaction", "field": "chain_id", "operator": "in", "value": ["1", "8453", "10"]}
valueis in wei (string). 1 ETH ="1000000000000000000"chain_idis string (e.g."1"for mainnet,"8453"for Base)tois checksummed address
2. ethereum_calldata
For decoded smart contract calls. Requires an abi field.
{
"field_source": "ethereum_calldata",
"field": "transfer.to",
"operator": "eq",
"value": "0xRecipient...",
"abi": {
"type": "function",
"name": "transfer",
"inputs": [
{"name": "to", "type": "address"},
{"name": "amount", "type": "uint256"}
]
}
}
Field format: <functionName>.<paramName> — references decoded parameter.
3. ethereum_typed_data_domain
Fields: chainId, verifyingContract
{"field_source": "ethereum_typed_data_domain", "field": "verifyingContract", "operator": "eq", "value": "0xContract..."}
{"field_source": "ethereum_typed_data_domain", "field": "chainId", "operator": "eq", "value": "1"}
4. ethereum_typed_data_message
For EIP-712 message fields. Requires a typed_data descriptor.
{
"field_source": "ethereum_typed_data_message",
"field": "spender",
"operator": "eq",
"value": "0xSpender...",
"typed_data": {
"types": {
"Permit": [
{"name": "owner", "type": "address"},
{"name": "spender", "type": "address"},
{"name": "value", "type": "uint256"}
]
},
"primary_type": "Permit"
}
}
5. ethereum_7702_authorization
Field: contract
{"field_source": "ethereum_7702_authorization", "field": "contract", "operator": "in", "value": ["0xA...", "0xB..."]}
6. solana_program_instruction
Field: programId
{"field_source": "solana_program_instruction", "field": "programId", "operator": "eq", "value": "11111111111111111111111111111111"}
{"field_source": "solana_program_instruction", "field": "programId", "operator": "in", "value": ["11111111111111111111111111111111", "TokenkegQfeZyiNwAJbNbGKPFXCWuBvf9Ss623VQ5DA"]}
7. solana_system_program_instruction
Fields: instructionName, Transfer.from, Transfer.to, Transfer.lamports
{"field_source": "solana_system_program_instruction", "field": "instructionName", "operator": "eq", "value": "Transfer"}
{"field_source": "solana_system_program_instruction", "field": "Transfer.to", "operator": "eq", "value": "RecipientPubkey..."}
{"field_source": "solana_system_program_instruction", "field": "Transfer.lamports", "operator": "lte", "value": "1000000000"}
lamportsis string. 1 SOL ="1000000000"(10^9)
8. solana_token_program_instruction
Fields: instructionName, TransferChecked.source, TransferChecked.destination, TransferChecked.authority, TransferChecked.amount, TransferChecked.mint
{"field_source": "solana_token_program_instruction", "field": "instructionName", "operator": "eq", "value": "TransferChecked"}
{"field_source": "solana_token_program_instruction", "field": "TransferChecked.mint", "operator": "eq", "value": "EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v"}
{"field_source": "solana_token_program_instruction", "field": "TransferChecked.amount", "operator": "lte", "value": "1000000"}
9. system
Field: current_unix_timestamp
{"field_source": "system", "field": "current_unix_timestamp", "operator": "lte", "value": "1735689600"}
Use for time-bounded policies (e.g. "allow transfers until 2025-01-01").
Common Policy Recipes
Use these as building blocks. Combine multiple conditions in one rule for AND logic. Use separate rules for OR logic.
EVM: Address Allowlist
Allow sending only to specific addresses:
{
"name": "Allowlist recipients",
"method": "eth_sendTransaction",
"conditions": [
{"field_source": "ethereum_transaction", "field": "to", "operator": "in", "value": ["0xAddr1...", "0xAddr2..."]}
],
"action": "ALLOW"
}
EVM: Transfer Value Cap
Allow transfers up to 0.1 ETH:
{
"name": "Max 0.1 ETH per tx",
"method": "eth_sendTransaction",
"conditions": [
{"field_source": "ethereum_transaction", "field": "value", "operator": "lte", "value": "100000000000000000"}
],
"action": "ALLOW"
}
EVM: Chain Restriction
Allow only on Base and Ethereum mainnet:
{
"name": "Base and mainnet only",
"method": "eth_sendTransaction",
"conditions": [
{"field_source": "ethereum_transaction", "field": "chain_id", "operator": "in", "value": ["1", "8453"]}
],
"action": "ALLOW"
}
EVM: Combined — Allowlist + Cap + Chain
{
"name": "Treasury transfers on Base, max 1 ETH",
"method": "eth_sendTransaction",
"conditions": [
{"field_source": "ethereum_transaction", "field": "to", "operator": "eq", "value": "0xTreasury..."},
{"field_source": "ethereum_transaction", "field": "value", "operator": "lte", "value": "1000000000000000000"},
{"field_source": "ethereum_transaction", "field": "chain_id", "operator": "eq", "value": "8453"}
],
"action": "ALLOW"
}
Allow All Operations (including message signing)
personal_sign and signMessage are not valid policy methods. Use * wildcard to allow them. Combine with specific DENY rules to restrict dangerous operations.
{
"name": "Allow all operations",
"method": "*",
"conditions": [],
"action": "ALLOW"
}
Typical pattern: DENY dangerous methods first, then ALLOW * for the rest:
[
{"name": "Block key export", "method": "exportPrivateKey", "conditions": [], "action": "DENY"},
{"name": "Allow everything else", "method": "*", "conditions": [], "action": "ALLOW"}
]
EVM: USDC Contract on Base
{
"name": "Allow USDC on Base",
"method": "eth_sendTransaction",
"conditions": [
{"field_source": "ethereum_transaction", "field": "to", "operator": "eq", "value": "0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913"},
{"field_source": "ethereum_transaction", "field": "chain_id", "operator": "eq", "value": "8453"}
],
"action": "ALLOW"
}
EVM: Block Private Key Export
{
"name": "Never export private key",
"method": "exportPrivateKey",
"conditions": [],
"action": "DENY"
}
EVM: Time-Bounded Access
Allow transfers until a specific date:
{
"name": "Allow until 2025-06-01",
"method": "eth_sendTransaction",
"conditions": [
{"field_source": "system", "field": "current_unix_timestamp", "operator": "lte", "value": "1748736000"}
],
"action": "ALLOW"
}
Solana: SOL Transfer Allowlist
{
"name": "Allow SOL to treasury",
"method": "signAndSendTransaction",
"conditions": [
{"field_source": "solana_system_program_instruction", "field": "instructionName", "operator": "eq", "value": "Transfer"},
{"field_source": "solana_system_program_instruction", "field": "Transfer.to", "operator": "eq", "value": "TreasuryPubkey..."}
],
"action": "ALLOW"
}
Solana: SOL Transfer Cap
{
"name": "Max 1 SOL per tx",
"method": "signAndSendTransaction",
"conditions": [
{"field_source": "solana_system_program_instruction", "field": "instructionName", "operator": "eq", "value": "Transfer"},
{"field_source": "solana_system_program_instruction", "field": "Transfer.lamports", "operator": "lte", "value": "1000000000"}
],
"action": "ALLOW"
}
Solana: SPL Token (USDC) Allowlist
{
"name": "Allow USDC transfers to recipient",
"method": "signAndSendTransaction",
"conditions": [
{"field_source": "solana_token_program_instruction", "field": "instructionName", "operator": "eq", "value": "TransferChecked"},
{"field_source": "solana_token_program_instruction", "field": "TransferChecked.mint", "operator": "eq", "value": "EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v"},
{"field_source": "solana_token_program_instruction", "field": "TransferChecked.destination", "operator": "eq", "value": "RecipientATA..."}
],
"action": "ALLOW"
}
Solana: Program Allowlist
Only allow interactions with specific programs:
{
"name": "Allow System and Token programs only",
"method": "signAndSendTransaction",
"conditions": [
{"field_source": "solana_program_instruction", "field": "programId", "operator": "in", "value": [
"11111111111111111111111111111111",
"TokenkegQfeZyiNwAJbNbGKPFXCWuBvf9Ss623VQ5DA"
]}
],
"action": "ALLOW"
}
Custom Restrictive Patterns
Use these only when the user explicitly requests tighter restrictions. Adapt the contract address and chain_id to the user's needs.
Contract-Specific Pattern
[
{"name": "Allow <DAPP_NAME>", "method": "eth_sendTransaction", "conditions": [
{"field_source": "ethereum_transaction", "field": "to", "operator": "eq", "value": "<CONTRACT_ADDRESS>"},
{"field_source": "ethereum_transaction", "field": "chain_id", "operator": "eq", "value": "<CHAIN_ID>"}
], "action": "ALLOW"},
{"name": "Token approvals on <NETWORK>", "method": "eth_sendTransaction", "conditions": [
{"field_source": "ethereum_transaction", "field": "chain_id", "operator": "eq", "value": "<CHAIN_ID>"},
{"field_source": "ethereum_transaction", "field": "value", "operator": "eq", "value": "0"}
], "action": "ALLOW"},
{"name": "Deny key export", "method": "exportPrivateKey", "conditions": [], "action": "DENY"}
]
Chain-Restricted Pattern
[
{"name": "Allow tx on <NETWORK>", "method": "eth_sendTransaction", "conditions": [
{"field_source": "ethereum_transaction", "field": "chain_id", "operator": "eq", "value": "<CHAIN_ID>"}
], "action": "ALLOW"},
{"name": "Allow signing on <NETWORK>", "method": "eth_signTypedData_v4", "conditions": [
{"field_source": "ethereum_typed_data_domain", "field": "chainId", "operator": "eq", "value": "<CHAIN_ID>"}
], "action": "ALLOW"},
{"name": "Deny key export", "method": "exportPrivateKey", "conditions": [], "action": "DENY"}
]
Wei / Lamports Quick Reference
EVM (wei)
| Amount | Wei String |
|---|---|
| 0.001 ETH | "1000000000000000" |
| 0.01 ETH | "10000000000000000" |
| 0.1 ETH | "100000000000000000" |
| 1 ETH | "1000000000000000000" |
| 10 ETH | "10000000000000000000" |
Formula: wei = eth * 10^18
Solana (lamports)
| Amount | Lamports String |
|---|---|
| 0.001 SOL | "1000000" |
| 0.01 SOL | "10000000" |
| 0.1 SOL | "100000000" |
| 1 SOL | "1000000000" |
| 10 SOL | "10000000000" |
Formula: lamports = sol * 10^9
USDC (6 decimals — same on EVM and Solana)
| Amount | Raw String |
|---|---|
| 1 USDC | "1000000" |
| 100 USDC | "100000000" |
| 1000 USDC | "1000000000" |
Chain IDs (EVM)
| Chain | ID (string in conditions) |
|---|---|
| Ethereum Mainnet | "1" |
| Ethereum Sepolia | "11155111" |
| Base | "8453" |
| Optimism | "10" |
| Arbitrum One | "42161" |
| Polygon | "137" |
| BSC | "56" |
Common Solana Program IDs
| Program | ID |
|---|---|
| System Program | 11111111111111111111111111111111 |
| Token Program | TokenkegQfeZyiNwAJbNbGKPFXCWuBvf9Ss623VQ5DA |
| Token-2022 | TokenzQdBNbLqP5VEhdkAS6EPFLC1PHnBqCXEpPxuEb |
| Associated Token | ATokenGPvbdGVxr1b2hvZbsiqW5xWH25efTNsLJA8knL |
| USDC Mint | EPjFWdd5AufqSSqeM2qN1xzybapC8G4wEGGkZwyTDt1v |
Interaction Guidelines
- Ask about chain first — Determine if the user needs EVM, Solana, or both policies.
- Clarify addresses — If the user says "allow transfers to my friend", ask for the address.
- Confirm amounts — Convert user-friendly amounts (e.g. "1 ETH") to wei/lamports in the output.
- Explain trade-offs — If a policy is overly permissive, warn the user. If too restrictive, note what will be blocked.
- Combine smartly — Use multiple conditions per rule for AND, multiple rules for OR.
- Always include deny-export — Recommend a
DENY exportPrivateKeyrule unless the user explicitly needs key export. - Output valid JSON — Every
json:policyblock must be valid, parseable JSON. Double-check addresses and values. - Rule name ≤ 50 characters — Privy API enforces a 50-character limit on rule names. NEVER include full on-chain addresses in rule names. Use short descriptive names like
"Allowlist recipients","Max 0.1 ETH per tx", or"Allow SOL to treasury". If you need to reference an address, abbreviate it (e.g.0xba86...BE6E).
Do NOT
- Do NOT output policy rules as code blocks or plain text — always call
wallet_propose_policy - Do NOT add extra fields to rules (only: name, method, conditions, action)
- Do NOT add extra fields to conditions (only: field_source, field, operator, value, plus abi/typed_data when needed)
- Do NOT use Solana field_sources (solana_*) with chain_type="ethereum" or vice versa
- Do NOT use lowercase action values — always "ALLOW" or "DENY"
- Do NOT use
ethereum_transactionorethereum_calldatafield_sources witheth_signTypedData_v4— useethereum_typed_data_domainorethereum_typed_data_messageinstead - Do NOT use
ethereum_transactionfield_sources witheth_sign7702Authorization— useethereum_7702_authorizationinstead - Do NOT use custom restrictive rules for a service unless the user explicitly asks for tighter restrictions — default to the wildcard policy
FAQ
What does the wallet-policy skill do?
Generate Privy wallet policy rules from natural language. Use when the user wants to set up, modify, or review wallet security policies — transfer limits, address allowlists, method restrictions, time windows, etc.
How do I install wallet-policy?
Run npx skills add https://github.com/starchild-ai-agent/official-skills --skill wallet-policy in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/wallet-policy/ (personal) or .claude/skills/wallet-policy/ (project).
Which AI agents can use wallet-policy?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is wallet-policy safe to install?
Security partners have reviewed it: Gen Agent Trust Hub passed; Socket passed; Snyk flagged items to review; Runlayer passed. As with any skill, read the SKILL.md and bundled scripts before installing.
Who maintains wallet-policy?
It is published by starchild-ai-agent in starchild-ai-agent/official-skills. It has 4,061 recorded installs.
More from starchild-ai-agent/official-skills
coinglass
starchild-ai-agent/official-skills
Crypto derivatives data: funding rates, open interest, liquidations, long/short ratios. Use when researching perp markets or comparing ETF flows (e.g. BTC funding, ETH OI). NOTE: platform API key is Startup tier — liquidation heatmap, coins-markets detail, liquidation order book and Hyperliquid positions require higher plans (see Plan Tiers).
wallet
starchild-ai-agent/official-skills
Multi-chain wallet: EVM and Solana balances, transfers, signing, and policy. Use when checking balances, sending tokens, signing typed data, or proposing a wallet policy (e.g. send 10 USDC on Base, sign EIP-712, Solana balance).
hyperliquid
starchild-ai-agent/official-skills
Trade perp futures, spot, and RWA on Hyperliquid DEX with up to asset max leverage. Use when placing perp or spot orders, setting TP/SL, or moving funds on Hyperliquid (e.g. long BTC 5x, sell ETH, deposit USDC, set stop).
coingecko
starchild-ai-agent/official-skills
Crypto spot prices, OHLC charts, market discovery, and global stats. Use when looking up coin prices, market caps, trending coins, sector rankings, or comparing tokens (e.g. BTC price, ETH chart, top gainers).
starchild-ai-agent/official-skills
Twitter/X data: fetch tweets, search, user profiles, followers, replies, trends. Use for any x.com or twitter.com URL or lookup (e.g. summarize this tweet, recent posts by @vitalikbuterin, search $SOL minfaves:50).
skill-creator
starchild-ai-agent/official-skills
Scaffold new skills with valid frontmatter, directory layout, and a starter SKILL.md. Use when building a new reusable workflow or wrapping a new API (e.g. create a kalshi skill, scaffold an API helper, start a charting skill).
Popular ai & agents skills
google-agents-cli-publish
google/agents-cli
This skill should be used when the user wants to "publish an agent", "publish my ADK agent", "register an agent with Gemini", "publish to Gemini" (formerly Gemini Enterprise), or needs guidance on the agents-cli publish gemini command. Also use when the user wants to "manage agents in Agent Registry", "list/update/delete registered agents", or "register an MCP server". Covers ADK vs A2A registration modes, programmatic and interactive usage, flag reference, auto-detection from deployment metadata, Agent Registry fleet management, and troubleshooting. Part of the agents-cli skills suite. Do NOT use for deployment (use google-agents-cli-deploy).
google-agents-cli-eval
google/agents-cli
This skill should be used when the user wants to "run an evaluation", "evaluate my agent", "evaluate my ADK agent", "write an eval dataset", "analyze eval failures", "compare eval results", "optimize agent", or needs guidance on the Gemini platform eval methodology and the Quality Flywheel. Covers eval metrics, dataset schema, LLM-as-judge scoring, and common failure causes. Applies to any agents-cli project, whatever framework the agent is written in. Do NOT use for agent API code patterns (ADK: use google-agents-cli-adk-code), deployment (use google-agents-cli-deploy), or project scaffolding (use google-agents-cli-scaffold).
skill-creator
anthropics/skills
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
codex-pet
prime-skills/runcomfy-agent-skills
Codex Pet generator on RunComfy. Build a Codex-compatible Codex Pet spritesheet.webp + pet.json from a single reference image, drop it into ${CODEXHOME:-$HOME/.codex}/pets/ / and Codex picks it up as a custom Codex Pet next to the 8 built-ins. This skill produces the exact Codex Pet atlas Codex expects (1536x1872 PNG/WebP, 8 cols x 9 rows, 192x208 cells, 9 animation states — idle, running-right, running-left, waving, jumping, failed, waiting, running, review). Calls OpenAI GPT Image 2 edit ONCE via the local RunComfy CLI as runcomfy run openai/gpt-image-2/edit to produce a canonical Codex Pet pose, then assembles all 9 animation rows programmatically with ImageMagick micro-transforms — no Codex Pro, no $imagegen, no OPENAIAPIKEY required, only RUNCOMFYTOKEN. Triggers on "codex pet", "create codex pet", "make codex pet", "hatch codex pet", "/hatch image", "desktop pet codex", "codex pets", "spritesheet.webp", or any explicit ask to build a custom pet for OpenAI Codex.
full-output-enforcement
leonxlnx/taste-skill
Overrides default LLM truncation behavior. Enforces complete code generation, bans placeholder patterns, and handles token-limit splits cleanly. Apply to any task requiring exhaustive, unabridged output.
executing-plans
obra/superpowers
Use when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available