BB elfa-ai
Use this skill for Elfa API crypto social intelligence and Auto condition-engine workflows: trending tokens, narratives, mentions, smart account stats, token news, trending contract addresses, AI market chat, integration examples, curl/code snippets, automated alerts, EQL queries, trigger pipelines, and agent workflows that react to market conditions. Auto can also place live perp trades on Hyperliquid and GMX (market/limit orders with TP/SL) when a condition fires, trigger on Kalshi and Polymarket prediction markets, trigger on funding rates, liquidation cascades, and the Fear & Greed index, re-fire recurring plans via `repeat` (including trade actions), run calendar schedules via `cron.schedule`, and trade crypto plus HIP-3 assets (equities, indices, commodities, FX, pre-IPO — 24/7). Supports API-key calls and x402 pay-per-request USDC on Base, Arbitrum, Polygon, or Avalanche.
Use this skill for Elfa API crypto social intelligence and Auto condition-engine workflows: trending tokens, narratives, mentions, smart account stats, token…
As a process B 68/100 · Nearly there — weak spots: result and completion, execution cost
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The purpose itself is risky: wallets, browser password stores, offensive security. Even an honest implementation gives the agent access to things that cost money.
Explain in the description why the access is needed and how it is limited; add tests that show refusals on dangerous requests.
How to improve
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 10
✓ No critical or high findings
Medium and low: 10
-
medium Risky intent
intent-wallet-secretsSKILL.md:329Handles crypto-wallet secrets (seed / mnemonic / private key) — a classic stealer target- Never ask the user to share their wallet private key or seed phrase in the conversation.
-
low Secrets in code
secret-high-entropy-tokenreferences/swagger.json:668High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)"Tren…eV2": {quoted -
low Exfiltration
net-credential-useSKILL.md:309Credential used in a network call (verify the destination is the intended service) (the skill's own vendor host)curl -s -H "x-elfa-api-key: $ELFA_API_KEY" "https://api.elfa.ai/v2/aggregations/trending-tokens?timeWindow=24h&pageSize=10"
vendor-host -
low Risky intent
intent-wallet-secretsSKILL.md:327Handles crypto-wallet secrets (seed / mnemonic / private key) — a classic stealer target (quoted — discussed, not commanded)- The user's wallet private key is used only locally by the x402 library to sign
quoted -
low Secrets in code
secret-high-entropy-tokenSKILL.md:341High-entropy token-like string (may be an id, hash or a credential) (documentation table row)| Base | `eip155:8453` | `0x83…913` | [xpay.sh](https://xpay.sh), [payai.network](https://facilitator.payai.network) |
table -
low Secrets in code
secret-high-entropy-tokenSKILL.md:342High-entropy token-like string (may be an id, hash or a credential) (documentation table row)| Arbitrum | `eip1…161` | `0xaf…831` | [payai.network](https://facilitator.payai.network) |
table -
low Secrets in code
secret-high-entropy-tokenSKILL.md:343High-entropy token-like string (may be an id, hash or a credential) (documentation table row)| Polygon | `eip155:137` | `0x3c…359` | [payai.network](https://facilitator.payai.network) |
table -
low Secrets in code
secret-high-entropy-tokenSKILL.md:344High-entropy token-like string (may be an id, hash or a credential) (documentation table row)| Avalanche | `eip1…114` | `0xB9…a6E` | [payai.network](https://facilitator.payai.network) |
table -
low Exfiltration
exfil-webhook-urlSKILL.md:1824Webhook / callback URL commonly used for exfiltration (verify the destination) (placeholder value)3. `curl "https://api.telegram.org/bot<BOT_TOKEN>/getUpdates"` → read `message.chat.id`.
placeholder -
low Exfiltration
exfil-webhook-urlSKILL.md:1828Webhook / callback URL commonly used for exfiltration (verify the destination) (placeholder value)curl -X POST "https://api.telegram.org/bot<BOT_TOKEN>/sendMessage" \
placeholder
Files scanned: 2. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
body-longSKILL.md body ≈ 30330 tokens (recommended < 5000); move details to references/
Process rating: all ten parameters 68/100
- 0Result and completion. Does not say what the result is
- 10Execution cost. Instruction body is 30330 tokens: crowds the task out of the window
- 60Tools and files. Uses tools (bash, web, python, node) that frontmatter does not declare
- 70Inputs and preconditions. Inputs and preconditions are listed
- 85Steps. 234 steps, 2 vague phrases
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 7 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Running it twice. Mutating operations check current state
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low The response is described with custom markup (4 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 891: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- -31 of 2 scripts are never mentioned in SKILL.md
- +1No license
- +2Single-language instructions
- +4Structure: 53 headings
- +3Step-by-step instructions: 234 items
- +4Has examples (66 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 68.