Skip to main content
Glama

Analyze feeds

analyze_feeds
Read-onlyIdempotent

Analyze feed health and explicit engagement to surface pruning candidates and estimate wasted posting volume using category-pooled Beta-Binomial bounds and optional volume sampling.

Instructions

Analyze feed health and explicit engagement with a category-pooled Beta-Binomial model (default 12 months, prior strength 10). By default counts starred, liked, broadcast and custom tags, excluding folders and passive system states; engagement_streams selects an explicit positive allowlist. Ordinary read flags are not engagement. Articles are deduplicated across positive streams. Categories with fewer than two engaged feeds use the global prior; multiple actual topic priors are averaged, excluding exclude_topic_folders. Reports exact equal-tailed 90% Beta bounds and posterior shapes only for complete, consistent counts. Missing, capped or failed coverage is unknown, not zero engagement. Retained engagement can exceed retained feed volume: this is flagged without silently flooring the denominator. Historical rates are retention/subscription-affected context, never authorization for automatic actions. Engagement and volume follow continuations within max_requests and the known 25-read reserve. Default sort unengaged_per_month estimates wasted posting volume; skipped sorts ascending by total minus engaged. Sorting engaged_count, title or days_since_newest skips engaged-feed volume reads; volume_sample optionally counts never-engaged active feeds. Costs 3 metadata reads plus engagement pages and optional volume pages EVERY call; no caching. max_requests counts base calls including failed calls; safe-read retries are extra (up to 4 attempts, counted in rate-limit status). Check get_rate_limit_status before large scans. See inoreader://usage for limits and URI examples.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "analyze_feeds"
  }
}

Input Schema

NameRequiredDescriptionDefault
limitNoMax number of feeds to return (default 100). Use to keep response size manageable.
folderNoLimit analysis to feeds in this folder
monthsNoTime window in months to analyze (default 12)
refreshNoNo-op, retained for compatibility. Reads are not cached between calls.
sort_byNoSort results by (default: unengaged_per_month). skipped sorts ascending by total - engaged to find feeds you read almost everything from (Must Read candidates). unengaged_per_month ranks by estimated wasted articles per month to surface pruning candidates. Use credible_lower for conservative ranking that penalizes small samples more.
max_requestsNoMaximum base Zone 1 calls across this workflow (default 100); retries are extra. Leaves 25 known requests in reserve and reports partial coverage.
volume_sampleNoHow many never-engaged active feeds to also count volume for, to rank noise (default 0). Set higher to find the noisiest unengaged feeds, at 1 extra API call each.
prior_strengthNoBeta prior pseudo-observation count (default 10). Higher values shrink small-sample feeds more aggressively toward the global mean.
engagement_pagesNoPer-tag safety cap on engagement pages (100 items per page, default 100). Pagination normally ends naturally when the ot time filter is exhausted; this cap only binds for users with thousands of engaged items per tag in the window.
engagement_streamsNoExplicit positive stream IDs. Only these streams count as engagement; omit to retain the default explicit-action/custom-tag policy. Do not include negative or automation labels.
exclude_topic_foldersNoFolder names or IDs excluded from category pooling (for example an operational priority folder); memberships are preserved.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed12 schema fields changedv1.6.1
    • changedInput schema / $schema
      Previous value: -"http://json-schema.org/draft-07/schema#"New value: +"https://json-schema.org/draft/2020-12/schema"
    • addedInput schema / properties / engagement_pages
      Added value: +{
      +  "description": "Per-tag safety cap on engagement pages (100 items per page, default 100). Pagination normally ends naturally when the ot time filter is exhausted; this cap only binds for users with thousands of engaged items per tag in the window.",
      +  "maximum": 500,
      +  "minimum": 1,
      +  "type": "integer"
      +}
    • addedInput schema / properties / engagement_streams
      Added value: +{
      +  "description": "Explicit positive stream IDs. Only these streams count as engagement; omit to retain the default explicit-action/custom-tag policy. Do not include negative or automation labels.",
      +  "items": {
      +    "minLength": 1,
      +    "type": "string"
      +  },
      +  "type": "array"
      +}
    • addedInput schema / properties / exclude_topic_folders
      Added value: +{
      +  "description": "Folder names or IDs excluded from category pooling (for example an operational priority folder); memberships are preserved.",
      +  "items": {
      +    "type": "string"
      +  },
      +  "type": "array"
      +}
    • changedInput schema / properties / limit / type
      Previous value: -"number"New value: +"integer"
    • addedInput schema / properties / max_requests
      Added value: +{
      +  "description": "Maximum base Zone 1 calls across this workflow (default 100); retries are extra. Leaves 25 known requests in reserve and reports partial coverage.",
      +  "maximum": 500,
      +  "minimum": 3,
      +  "type": "integer"
      +}
    • changedInput schema / properties / months / description
      Previous value: -"Time window in months to analyze (default 3)"New value: +"Time window in months to analyze (default 12)"
    • changedInput schema / properties / refresh / description
      Previous value: -"Force fresh data by clearing the cache before running (default false)"New value: +"No-op, retained for compatibility. Reads are not cached between calls."
    • changedInput schema / properties / sort_by / description
      Previous value: -"Sort results by (default: engagement_rate). Use credible_lower for conservative ranking that penalizes small samples more."New value: +"Sort results by (default: unengaged_per_month). skipped sorts ascending by total - engaged to find feeds you read almost everything from (Must Read candidates). unengaged_per_month ranks by estimated wasted articles per month to surface pruning candidates. Use credible_lower for conservative ranking that penalizes small samples more."
    • changedInput schema / properties / sort_by / enum
      Previous value: -[
      -  "engagement_rate",
      -  "credible_lower",
      -  "saved_count",
      -  "title",
      -  "days_since_newest"
      -]New value: +[
      +  "engagement_rate",
      +  "credible_lower",
      +  "engaged_count",
      +  "skipped",
      +  "title",
      +  "days_since_newest",
      +  "unengaged_per_month"
      +]
    • removedInput schema / properties / starred_pages
      Removed value: -{
      -  "description": "Max pages of saved/starred articles to fetch (100 per page, default 10). More pages = better engagement data but higher API cost.",
      -  "maximum": 20,
      -  "minimum": 1,
      -  "type": "number"
      -}
    • addedInput schema / properties / volume_sample
      Added value: +{
      +  "description": "How many never-engaged active feeds to also count volume for, to rank noise (default 0). Set higher to find the noisiest unengaged feeds, at 1 extra API call each.",
      +  "maximum": 200,
      +  "minimum": 0,
      +  "type": "integer"
      +}
  2. First observedv0.1.0

TDQS

A4.1/5.0
Behavior5/5

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

Annotations already declare read-only, idempotent, non-destructive behavior, and the description goes well beyond that: cost model (3 metadata reads plus engagement pages EVERY call, no caching), retry/reserve accounting, the 25-read reserve, and the crucial semantic that missing/capped/failed coverage is 'unknown, not zero'. It even flags that retained engagement can exceed retained volume rather than silently flooring it, and that historical rates are not authorization for actions.

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?

Nearly every sentence carries information, but the wall of dense jargon ('equal-tailed 90% Beta bounds', 'posterior shapes', 'category-pooled Beta-Binomial') makes it hard to parse on first read. Front-loading is partial and the ordering jumps between model semantics, cost, sorting, and rate limits, hurting scanability.

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?

For an 11-parameter, cost-bearing analytical tool with no output schema, the description covers cost model, coverage/unknown semantics, sort behavior, and partial return characteristics (Beta bounds, posterior shapes). It could say more about exact return shape and partial-coverage reporting, but an agent has enough to invoke it correctly.

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

Parameters4/5

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

With 100% schema coverage the baseline is 3, but the description adds meaning the schema lacks: engagement_streams is an explicit positive allowlist, ordinary read flags do not count, articles deduplicate across streams, and certain sorts skip volume reads while volume_sample counts never-engaged feeds. It enriches several parameters beyond their schema text without fully restating every one.

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

Purpose4/5

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

Opens with a specific verb and resource ('Analyze feed health and explicit engagement') and names the underlying model and default window, which distinguishes it from siblings like get_unread_counts and suggest_feed_cleanup. Sibling differentiation is present but scattered (pruning candidates, Must Read candidates) rather than stated up front, keeping it below a 5.

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

Usage Guidelines4/5

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

Rich context: it tells the agent what each sort is for (unengaged_per_month for pruning candidates, skipped for Must Read candidates, credible_lower for conservative ranking) and warns to check get_rate_limit_status before large scans. It stops short of explicit when-not-to-use-a-sibling statements, so it is clear guidance without hard exclusions.

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