agentleFS
aeonfunaeonfun/aeon/skills/flovia/SKILL.md

flovia

Skill789 starsChanged 23 days agoMITView on GitHub

Live lookups through the Flovia MCP - Nansen smart money and token flows, CoinGecko markets, direct chain reads, Zerion wallets, Exa web search and Firecrawl page reads - on a free shared allowance. OAuth Connect or an flv_ key in the dashboard MCP panel.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Ask nothing: No consent question, no "pick a next step" prompt - next_options in a reply are listed in the notify as suggestions, never …

    line 39
  2. Very specific:

    Start with flovia_status (allowance %, resets, service states, recent_updates).

    line 48
  3. Very specific:

    Run the planned calls, one question at a time.

    line 60

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Flovia

${var} - the lookup question, e.g. is token 0x... on Base being accumulated by smart money or what does wallet 0x... hold across chains. Empty = a short status report (free allowance, services, recent changes). Prefer plain language; don't name endpoints.

Answer one question with live data through the Flovia MCP (api.flovia.dev/mcp): Nansen smart money / holders / token flows / wallet profiles, CoinGecko prices and DEX pools, direct chain reads, Zerion wallets, Exa public web search and Firecrawl public-page reads, plus optional OpenAI text generation. Every metered call draws from the operator's free allowance (or purchased balance), so calls are deliberate.

Detection & auth

The server is wired from the dashboard MCP panel: one-click Connect (OAuth) or an flv_ API key pasted as MCP_FLOVIA_TOKEN. Its tools surface as mcp__flovia__* - discover them from the server; tool descriptions are the source of truth. Flovia is in alpha: call flovia_status first and read recent_updates - tools, limits and prices may have changed since this file was written.

  • No mcp__flovia__* tool callable → the server isn't connected. Log FLOVIA_NOT_CONNECTED, notify once pointing the operator at the dashboard → MCP → Connect Flovia, and exit.
  • 401 / invalid token → the OAuth refresh failed (see docs/mcp-oauth.md) or the key was revoked. Log FLOVIA_AUTH_STALE, notify the operator to reconnect once, and exit.
  • A refusal reply carries explanation (by, code, executed, meaning, next_step). Follow next_step; FLOVIA_WINDOW_FULL when a free window is exhausted (the reply says when it resets).

Aeon runs (scheduled or dispatched)

An Aeon run has no user to talk to:

  • Ask nothing. No consent question, no "pick a next step" prompt - next_options in a reply are listed in the notify as suggestions, never executed.
  • Read-only. Flovia can prepare unsigned swaps and Hyperliquid actions; never call those tools. Read tools only (flovia_nansen, flovia_coingecko, flovia_quicknode, flovia_zerion, flovia_exa, flovia_firecrawl, flovia_status, flovia_usage_today, flovia_lookup, flovia_services list). flovia_openai/flovia_minimax/flovia_fal/flovia_runway are generation tools - off by default; use only if ${var} explicitly asks.
  • Stable operation ids. Pass one operation_id per question (flovia-<date>-<short-slug>). If a reply is lost or executed is unknown, resend with the same id - it is never charged twice. Never retry under a new id, and never auto-retry a pending/unknown result.
  • Never put secrets in lookups. Request text and purposes are kept by Flovia for up to 2 years during the alpha, and Nansen lookups are shared with Nansen. Keep request text to public facts; no keys, seeds, credentials, or the operator's private info. This is the data-use terms the operator accepted when registering - mention it once in the notify when Nansen data is used.

Steps

1. Plan

Start with flovia_status (allowance %, resets, service states, recent_updates). Then map ${var} to sources:

  • Prices, market caps, charts, DEX pools → flovia_coingecko
  • A public wallet's portfolio, positions, PnL, transactions → flovia_zerion
  • A transaction's status/gas, a balance, a contract read, event logs → flovia_quicknode (chain: "list" for the configured chains)
  • Smart money, holders, token flows, wallet labels → flovia_endpoints to pick the Nansen endpoint, then flovia_nansen
  • Public web research / cited sources → flovia_exa; one public page or PDF → flovia_firecrawl

Spend discipline: CoinGecko, Quicknode, Firecrawl and Zerion are flat-rate (separate daily limits, don't touch the free allowance). Nansen, Exa, context.dev and generation are metered at a much higher cost. Prefer flat-rate sources; use Nansen only when the question needs smart-money data - at most 2 Nansen calls per run, ≤ 8 tool calls total. When a free window is past 75%, stop at the flat-rate answer and say so.

2. Fetch

Run the planned calls, one question at a time. For a repeated check (${var} like watch: <question>), compare against the previous run at output/.chains/flovia.md and report what changed. Reopening an earlier answer via flovia_lookup is free - use it instead of re-asking.

3. Answer

Trace every figure to a tool response; cite source URLs for Exa results. Distinguish observed fact from inference; keep units, dates and gaps.

4. Notify

Deliver via ./notify -f <file>: the answer, the evidence, allowance state if a window crossed 75%, and next_options as text suggestions. Exactly one ./notify call per run - it overwrites $AEON_PENDING_DIR/.pending-flovia.md which becomes the chain artifact output/.chains/flovia.md.

5. Result record

This skill is read-only, so the workflow's read-only guard writes its ### flovia log entry from your captured output; a self-written entry would be a duplicate. Put this record in your final output:

### flovia
- Query: <${var}, truncated>
- Result: FLOVIA_OK | FLOVIA_EMPTY_VAR | FLOVIA_NOT_CONNECTED | FLOVIA_AUTH_STALE | FLOVIA_WINDOW_FULL | FLOVIA_ERROR
- Sources: <tool names used> | Calls: <N>/<budget>

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.