# io.github.cyanheads/first-epss-mcp-server (remote · first-epss.caseyjhand.com)

Look up FIRST EPSS scores and daily history, find CVEs by threshold, and prioritize them with CVSS.

- Trust score: 80/100 (high trust)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-05

## Components

- remote · `first-epss.caseyjhand.com`: 80/100 (this document), [markdown](https://verifymcp.io/servers/cyanheads-first-epss-mcp-server/first-epss.md), [page](https://verifymcp.io/servers/cyanheads-first-epss-mcp-server/first-epss)
- npm · `@cyanheads/first-epss-mcp-server`: 38/100, [markdown](https://verifymcp.io/servers/cyanheads-first-epss-mcp-server/cyanheads-first-epss-mcp-server.md), [page](https://verifymcp.io/servers/cyanheads-first-epss-mcp-server/cyanheads-first-epss-mcp-server)

## Channel facts

- Endpoint: `https://first-epss.caseyjhand.com/mcp`
- Transports: `streamable-http`
- Auth: `none`
- Version: `0.2.0`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-05.

- **Endpoint Security**: 83/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - No authorisation is required to call this server. Every tool declares its destructiveHint and none is destructive, so open access doesn't expose one.
  - HTTPS is enforced; there's no plaintext access path.
  - The HSTS (Strict-Transport-Security) header is present.
  - DNSSEC is configured correctly; the domain's records validate against the full chain to the root.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 73/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 1449 tokens (~241/item across 6 items; 5 tools + 1 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 40/100
  - Stability observed for 12 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 5 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 7 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a current MCP spec version (2026-07-28).

## Install

### How do I install the io.github.cyanheads/first-epss-mcp-server server?

io.github.cyanheads/first-epss-mcp-server is a hosted endpoint at https://first-epss.caseyjhand.com/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add --transport http cyanheads-first-epss-mcp-server 'https://first-epss.caseyjhand.com/mcp'
```

### Cursor

```json
{
  "mcpServers": {
    "cyanheads-first-epss-mcp-server": {
      "url": "https://first-epss.caseyjhand.com/mcp"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "cyanheads-first-epss-mcp-server": {
      "type": "http",
      "url": "https://first-epss.caseyjhand.com/mcp"
    }
  }
}
```

### Codex

```toml
[mcp_servers.cyanheads-first-epss-mcp-server]
url = "https://first-epss.caseyjhand.com/mcp"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "cyanheads-first-epss-mcp-server": {
      "type": "remote",
      "url": "https://first-epss.caseyjhand.com/mcp",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add cyanheads-first-epss-mcp-server --url 'https://first-epss.caseyjhand.com/mcp' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  cyanheads-first-epss-mcp-server:
    url: "https://first-epss.caseyjhand.com/mcp"
```

### Netclaw

```json
{
  "McpServers": {
    "cyanheads-first-epss-mcp-server": {
      "Transport": "http",
      "Url": "https://first-epss.caseyjhand.com/mcp"
    }
  }
}
```

### Vellum

```bash
assistant mcp add cyanheads-first-epss-mcp-server -t streamable-http -u 'https://first-epss.caseyjhand.com/mcp'
```

### Other

```json
{
  "mcpServers": {
    "cyanheads-first-epss-mcp-server": {
      "type": "http",
      "url": "https://first-epss.caseyjhand.com/mcp"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-04 (score 80, +1)

No change was recorded against any check on this day. Stability & Change Management went from 33 to 37. That category is still filling its 30-day observation window: 10 days of observed history at the previous scan, 11 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-02 (score 79, +1)

No change was recorded against any check on this day. Stability & Change Management went from 27 to 30. That category is still filling its 30-day observation window: 8 days of observed history at the previous scan, 9 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-31 (score 78, +1)

No change was recorded against any check on this day. Stability & Change Management went from 20 to 23. That category is still filling its 30-day observation window: 6 days of observed history at the previous scan, 7 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-29 (score 77, +4)

- [security] Tool “epss_query_threshold” rewrote its description, which is the text the model reads
- [functional] First check of Schema quality: 100
- [functional] New resource “first-epss-guide”
- [functional] Server version: 0.1.3 → 0.2.0

### 2026-10-27 (score 73, +1)

No change was recorded against any check on this day. Stability & Change Management went from 7 to 10. That category is still filling its 30-day observation window: 2 days of observed history at the previous scan, 3 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-25 (score 72, +1)

- [functional improvement] Stability: unverified → 0.03

### 2026-10-24 (score 71)

First indexed and scored.

## MCP tools (5)

### `epss_get_scores` (~199 tokens)

Epss Get Scores

Retrieve latest available or historically dated FIRST EPSS probability and percentile for supplied CVEs. Process up to 500 unique entries from batchOffset; response byte limits can return fewer whole entries. Resume with nextBatchOffset as batchOffset, the original cves and returned score date as date. deferredCves were not queried; withheldCves were processed but omitted, and either list can be a declared preview. Missing scores remain unknown; use epss_list_reference for coverage and interpretation.

Input parameters:

- `batchOffset` (integer): Resume in first-occurrence normalized unique input order; resubmit the original cves, returned score date and original policy/severity inputs.
- `cves` (array, required): One to 5000 CVE identifiers; whitespace, comma or semicolon-delimited strings are also accepted.
- `date` (string): UTC date YYYY-MM-DD from 2021-04-14 through today; omitted or blank selects latest available data.

Output parameters:

- `batchOffset` (integer): Starting offset in the original normalized unique input.
- `budgetLimited` (boolean): True when whole entries or accounting lists were shortened to fit the serialized response budget.
- `deferredCount` (integer): Identifiers beyond this call workload cap, not queried.
- `deferredCves` (array): Never-queried identifiers beyond this call workload cap; complete unless deferredCvesComplete is false. Resume using nextBatchOffset and the original submission.
- `deferredCvesComplete` (boolean): True when deferredCves includes every never-queried deferred identifier.
- `duplicateCount` (integer): Duplicate submitted entries removed.
- `error` (object): Present when the call failed. Absent on success.
- `nextBatchOffset` (integer): Resume with this batchOffset and the original cves, returned date and original severity/policy inputs.
- `notice` (string): Missing-data, deferred-work or model-metadata caveats.
- `processedCount` (integer): Unique identifiers processed, including invalid formats.
- `provenance` (object): Source and model attribution.
- `requestedCount` (integer): Unique normalized identifiers submitted.
- `returnedCount` (integer): Processed identifiers whose complete score, history or unscored entry is included.
- `scores` (array): Forecasts ordered by EPSS descending then CVE.
- `unscoredCves` (array): Retained processed identifiers without scores, in request order.
- `withheldCount` (integer): Processed identifiers withheld by the response byte budget; queried valid entries are included in processedCount.
- `withheldCves` (array): Already processed identifiers withheld by the byte budget, in original input order; may be a preview.
- `withheldCvesComplete` (boolean): True when withheldCves includes every withheld identifier.

### `epss_get_timeseries` (~226 tokens)

Epss Get Timeseries

Retrieve an EPSS anchor and up to 30 preceding daily observations for up to 50 unique CVEs from batchOffset. Response byte limits retain whole history items. Resume CVE batches with nextBatchOffset as batchOffset, the original cves and returned score date as date; withheldCves were processed, deferredCves were not queried, and lists can be declared previews. Exact 7-day and 30-day deltas require their calendar baselines; gaps remain unknown. Model boundaries do not explain score movement. Re-anchor using previousAnchorDate for the preceding history window; that snapshot may be unavailable.

Input parameters:

- `batchOffset` (integer): Resume in first-occurrence normalized unique input order; resubmit the original cves, returned score date and original policy/severity inputs.
- `cves` (array, required): One to 5000 CVE identifiers; whitespace, comma or semicolon-delimited strings are also accepted.
- `date` (string): UTC date YYYY-MM-DD from 2021-04-14 through today; omitted or blank selects latest available data.

Output parameters:

- `batchOffset` (integer): Starting offset in the original normalized unique input.
- `budgetLimited` (boolean): True when whole entries or accounting lists were shortened to fit the serialized response budget.
- `deferredCount` (integer): Identifiers beyond this call workload cap, not queried.
- `deferredCves` (array): Never-queried identifiers beyond this call workload cap; complete unless deferredCvesComplete is false. Resume using nextBatchOffset and the original submission.
- `deferredCvesComplete` (boolean): True when deferredCves includes every never-queried deferred identifier.
- `duplicateCount` (integer): Duplicate submitted entries removed.
- `error` (object): Present when the call failed. Absent on success.
- `items` (array): Derived trajectories ordered by anchor EPSS descending then CVE.
- `nextBatchOffset` (integer): Resume with this batchOffset and the original cves, returned date and original severity/policy inputs.
- `notice` (string): Missing-data, deferred-work or model-metadata caveats.
- `processedCount` (integer): Unique identifiers processed, including invalid formats.
- `provenance` (object): Source and model attribution.
- `requestedCount` (integer): Unique normalized identifiers submitted.
- `returnedCount` (integer): Processed identifiers whose complete score, history or unscored entry is included.
- `unscoredCves` (array): Retained processed identifiers without scores, in request order.
- `withheldCount` (integer): Processed identifiers withheld by the response byte budget; queried valid entries are included in processedCount.
- `withheldCves` (array): Already processed identifiers withheld by the byte budget, in original input order; may be a preview.
- `withheldCvesComplete` (boolean): True when withheldCves includes every withheld identifier.

### `epss_query_threshold` (~326 tokens)

Epss Query Threshold

Discover CVEs using inclusive EPSS or percentile bounds, CVE year or dataset-addition recency; filters intersect and at least one is required. Examine one candidate page in the upstream FIRST order, which is not ranked by EPSS, so the first page does not hold the highest scores. Each serialized response surface fits 65536 UTF-8 bytes; budgetLimited and withheldCount disclose whole matches withheld, with nextOffset at the first withheld match. Year filtering can return an empty page with nextOffset; candidateTotal is a broad-query total, not an exact-year total. Continue with the same filters, limit, returned score date and nextOffset.

Input parameters:

- `addedWithinDays` (integer): Dataset-addition recency relative to upstream present, independent of historical date.
- `date` (string): UTC date YYYY-MM-DD from 2021-04-14 through today; omitted or blank selects latest available data.
- `limit` (integer): Requested candidates per page; positive safe integers above 10000 are clamped to 10000 before exact-year filtering.
- `maxEpss` (number): Inclusive maximum EPSS probability.
- `maxPercentile` (number): Inclusive maximum population percentile rank.
- `minEpss` (number): Inclusive minimum EPSS probability.
- `minPercentile` (number): Inclusive minimum population percentile rank.
- `offset` (integer): Candidate-space offset; use returned nextOffset for continuation.
- `year` (integer): Exact CVE year from 1999 through the current UTC year; candidate pages may contain no matches.

Output parameters:

- `error` (object): Present when the call failed. Absent on success.
- `filters` (object): Applied filters; combine with provenance scoreDates to pin continuation.
- `notice` (string): Missing-data, deferred-work or model-metadata caveats.
- `pagination` (object): Candidate pagination accounting.
- `provenance` (object): Snapshot source and model metadata, including dates from filtered-out candidates.
- `scores` (array): Retained records from this candidate page.

### `epss_prioritize_cves` (~279 tokens)

Epss Prioritize Cves

Assign remediation tiers using latest or dated EPSS and caller-supplied CVSS base severity. Defaults are severity >= 7 and EPSS probability >= 0.10; select percentile explicitly for population rank. Missing severity stays in epssOnly and missing EPSS stays unscored. Thresholds are policy conventions, not FIRST guidance. Process up to 500 unique entries from batchOffset, returning whole entries within the byte budget. Resume using nextBatchOffset as batchOffset, the original cves with original cvssBaseScore values, returned score date as date, and the same severityThreshold and likelihood. Deferred lists may be declared previews; withheldCves were already processed.

Input parameters:

- `batchOffset` (integer): Resume in first-occurrence normalized unique input order; resubmit the original cves, returned score date and original policy/severity inputs.
- `cves` (array, required): One to 5000 CVEs with optional caller severity; process up to 500 unique identifiers from batchOffset.
- `date` (string): UTC date YYYY-MM-DD from 2021-04-14 through today; omitted or blank selects latest available data.
- `likelihood`: Select one likelihood metric and threshold; omitted uses EPSS >= 0.10.
- `severityThreshold` (number): Inclusive high severity threshold, default 7.

Output parameters:

- `batchOffset` (integer): Starting offset in the original normalized unique input.
- `budgetLimited` (boolean): True when whole entries or accounting lists were shortened to fit the serialized response budget.
- `deferredCount` (integer): Identifiers beyond this call workload cap, not queried.
- `deferredCves` (array): Never-queried identifiers beyond this call workload cap; complete unless deferredCvesComplete is false. Resume using nextBatchOffset and the original submission.
- `deferredCvesComplete` (boolean): True when deferredCves includes every never-queried deferred identifier.
- `duplicateCount` (integer): Duplicate submitted entries removed.
- `epssOnly` (array): Scored CVEs without caller severity; never classified as low severity.
- `error` (object): Present when the call failed. Absent on success.
- `nextBatchOffset` (integer): Resume with this batchOffset and the original cves, returned date and original severity/policy inputs.
- `notice` (string): Missing-data, deferred-work or model-metadata caveats.
- `processedCount` (integer): Unique identifiers processed, including invalid formats.
- `provenance` (object): Source and model attribution.
- `requestedCount` (integer): Unique normalized identifiers submitted.
- `returnedCount` (integer): Processed identifiers whose complete score, history or unscored entry is included.
- `thresholds` (object): Applied policy conventions, not FIRST recommendations.
- `tiers` (object): Policy tiers, each sorted by EPSS, CVSS descending then CVE ascending.
- `unscoredCves` (array): Processed identifiers without EPSS, preserving supplied severity.
- `withheldCount` (integer): Processed identifiers withheld by the response byte budget; queried valid entries are included in processedCount.
- `withheldCves` (array): Already processed identifiers withheld by the byte budget, in original input order; may be a preview.
- `withheldCvesComplete` (boolean): True when withheldCves includes every withheld identifier.

### `epss_list_reference` (~79 tokens)

Epss List Reference

Explain FIRST EPSS interpretation, model publishing dates, lookup coverage and transparent prioritization policy. Use this static reference before interpreting missing scores, score movement or probability versus percentile thresholds.

Input parameters:

- `topic` (string): interpretation: score meaning; models: publishing history; coverage: limits and missing data; prioritization: tier policy; all: every topic.

Output parameters:

- `error` (object): Present when the call failed. Absent on success.
- `topics` (array): Selected reference topics.
- `verifiedAt` (string): Primary-source verification date.

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/cyanheads-first-epss-mcp-server/first-epss#diagnostics

## Score history

- 2026-11-05: 80
- 2026-11-04: 80
- 2026-11-03: 79
- 2026-11-02: 79
- 2026-11-01: 78
- 2026-10-31: 78
- 2026-10-30: 77
- 2026-10-29: 77
- 2026-10-28: 73
- 2026-10-27: 73
- 2026-10-26: 72
- 2026-10-25: 72
- 2026-10-24: 71

## Common questions

### What is the io.github.cyanheads/first-epss-mcp-server server?

io.github.cyanheads/first-epss-mcp-server is listed in the public MCP registry as io.github.cyanheads/first-epss-mcp-server. Look up FIRST EPSS scores and daily history, find CVEs by threshold, and prioritize them with CVSS. This page covers its hosted endpoint (https://first-epss.caseyjhand.com/mcp).

### Is the io.github.cyanheads/first-epss-mcp-server server safe to use?

io.github.cyanheads/first-epss-mcp-server scores 80 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the io.github.cyanheads/first-epss-mcp-server server expose?

io.github.cyanheads/first-epss-mcp-server exposes 5 tools: epss_get_scores, epss_get_timeseries, epss_query_threshold, epss_prioritize_cves, epss_list_reference. Their descriptions and schemas cost roughly 1,109 tokens of context every time the server is loaded.

### Does the io.github.cyanheads/first-epss-mcp-server server require authentication?

No. We connected to io.github.cyanheads/first-epss-mcp-server without credentials and it answered, so anything it exposes is reachable by anyone who knows the address.

### Is the io.github.cyanheads/first-epss-mcp-server server still maintained?

io.github.cyanheads/first-epss-mcp-server is still listed as active in the MCP registry. We last reached this channel on 5 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

## Links

- Remote endpoint: https://first-epss.caseyjhand.com/mcp
- Repository: https://github.com/cyanheads/first-epss-mcp-server
- Changelog RSS feed: https://verifymcp.io/servers/cyanheads-first-epss-mcp-server/first-epss.xml
- Changelog JSON feed: https://verifymcp.io/servers/cyanheads-first-epss-mcp-server/first-epss.json
- HTML version of this page: https://verifymcp.io/servers/cyanheads-first-epss-mcp-server/first-epss