Skip to main content
Glama

f1ow

by com.f1ow
OverviewDiscussionsAdmin

Positioning Analysis

research_positioning
Read-only

Deterministic positioning read for one coin — the game-state companion to research_regime, answering 'who is crowded, who is paying, who is trapped, and where'. Up to nine axes, each tagged with the market-actor cohort it reads: crowding (annualized funding vs the venue-standard 0.01%/8h anchor, long/short skew, dollars/day the majority pays to hold, crowd-vs-top-trader divergence, and cross-venue funding dispersion), buildup (24h price vs open-interest direction and the funding trend — whether positions are being added into the move or closed out), liquidity (order-book spread and imbalance — makers backing off), basis (perp vs oracle premium — spot-perp froth or hedging pressure, informational), fragility (24h liquidations as a share of market-wide open interest; the axis degrades rather than substituting a single venue's OI, since the liquidation total is cross-exchange), predation (visible Hyperliquid whale positions within 2 daily sigmas of their liquidation price, with defending book depth), trap (share of tracked whale notional underwater on the funding-paying side — attrition fuel), house book (Hyperliquid HLP LP-vault inventory, summed across its child sub-vaults — the literal venue counterparty, so the inventory it carries mirrors how traders are crowded; measured as net over its own gross book, since a market maker's net is a rounding error against venue open interest; informational), and disagreement (the most liquid matching prediction market, volume-gated, informational). Each axis reports its state, the numbers, and a +1/0/-1 fragility vote (+1 clean, -1 crowded/fragile); the stance is the disclosed sum-of-votes rule. Also returns a one-line 'farmed' synthesis (which cohort the board is currently farming, or an explicit statement that none is), focal points — reachable liquidation and breakeven price levels the whole market can see — and 'would change the call' thresholds a polling agent can watch statelessly. Reports the board, not a direction: crowding says who is paying, not where price goes. Deterministic — same inputs, same read. Call research_regime first for market context, then this per coin before sizing or timing decisions.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
symbolYesCoin symbol, e.g. "BTC" or "ETH".

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault
axesNo
farmedNoOne-line synthesis of which cohort the board is currently farming. Always populated: when no cohort clears the trap/crowding/predation bars it says so explicitly and names who is paying the carry.
stanceNoOverall positioning-fragility label
symbolNo
axesTotalNo
stanceRuleNoThe exact rule that produced stance
thresholdsNoThe wouldChange lines in machine-readable form for stateless alerting
focalPointsNoPrice levels that are common knowledge on the board, nearest to spot first
skippedAxesNoAxes attempted but not computed; the stance is over the computed ones only
wouldChangeNoThresholds that would flip an axis, as prose
axesComputedNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed1 schema field changed
    • addedOutput schema / properties / skippedAxes
      Added value: +{
+  "description": "Axes attempted but not computed; the stance is over the computed ones only",
+  "items": {
+    "properties": {
+      "name": {
+        "type": "string"
+      },
+      "reason": {
+        "description": "The feed that was offline, or what the coin lacks",
+        "type": "string"
+      }
+    },
+    "type": "object"
+  },
+  "type": "array"
+}
  2. First observed

TDQS

A4.2/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations give readOnlyHint=true, so safety is already covered; the description usefully adds determinism ('same inputs, same read'), the explicit stance rule (sum-of-votes), the degradation behavior of the fragility axis, and the critical framing 'Reports the board, not a direction — crowding says who is paying, not where price goes.' Much of the remaining text enumerates return content rather than behavior, which blunts it slightly.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with a strong orienting sentence, but the axis catalogue runs to roughly 350 words for a one-parameter tool, and a large share of that detail restates return values already carried by the output schema. Dense and hard to scan; a good chunk does not earn its place.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Nothing an agent needs to select or invoke the tool is missing: scope, sequencing, stance semantics, and determinism are all present. It undershoots on conciseness rather than completeness, over-explaining return values that the output schema already provides.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Single parameter with 100% schema description coverage, so the schema already documents 'symbol' fully. The description adds no syntax, format, or edge-case guidance for the parameter. Baseline 3 applies when the schema does the heavy lifting.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a specific verb+resource and scope ('Deterministic positioning read for one coin'), then frames it as 'the game-state companion to research_regime' and enumerates the exact question it answers ('who is crowded, who is paying, who is trapped, and where'). An agent can distinguish it from research_regime, funding_current, and flow_whale_positions without opening any schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit sequencing and context: 'Call research_regime first for market context, then this per coin before sizing or timing decisions.' It names the sibling to chain from and the decision point at which this tool belongs, leaving nothing to inference.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All f1ow toolsInspect
Try in Browser
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources