Skip to content

Add tensorfeed skill: live AI ecosystem data via hosted MCP#1114

Open
RipperMercs wants to merge 1 commit into
anthropics:mainfrom
RipperMercs:add-tensorfeed-skill
Open

Add tensorfeed skill: live AI ecosystem data via hosted MCP#1114
RipperMercs wants to merge 1 commit into
anthropics:mainfrom
RipperMercs:add-tensorfeed-skill

Conversation

@RipperMercs

Copy link
Copy Markdown

Summary

Adds the TensorFeed skill that teaches Claude when and how to use TensorFeed.ai's hosted MCP server (https://tensorfeed.ai/api/mcp) for live, machine-readable AI ecosystem data and adjacent regulated-domain data.

What it covers

The MCP server exposes 17 free tools, no auth required, across five domains:

Domain Tools Sources
News + AI ecosystem 3 12+ AI news sources, status pages, model catalog
Security advisories 5 MITRE CVE, CISA KEV, FIRST.org EPSS, OSV.dev
Finance + filings 3 SEC EDGAR (full-text search + submissions + tickers)
FDA regulatory 5 openFDA: FAERS, SPL, drug/food recalls, MAUDE
Energy + macro 1 EIA Open Data

The SKILL.md walks Claude through connection, trigger conditions, the tool catalog, and output discipline (preserve attribution, surface source/fetched_at for freshness).

Trigger guidance

The frontmatter description includes explicit TRIGGER and SKIP sections per the convention seen in skills/claude-api/SKILL.md:

  • Trigger: AI news, model pricing comparison, AI service status, security CVE lookup, FDA recall lookup, SEC company filing, energy/macro indicator
  • Skip: pure code generation without data dependency, user wants their own private data, non-AI-domain questions

License

  • Skill content: MIT
  • Underlying data: most is US Government public domain (SEC, BLS, FRED, MITRE, CISA, EIA) or CC0 (openFDA), with FIRST.org EPSS free for any use and OSV.dev under Apache 2.0. Commercial redistribution permitted across the surface; attribution preserved on every response per upstream policies.

Test plan

  • marketplace.json validates as JSON
  • SKILL.md frontmatter parses (name, description, license)
  • MCP server live and responding to tools/list (17 tools verified 2026-05-09)
  • Each tool returns 200 + ok:true on representative queries
  • /plugin install tensorfeed@anthropic-agent-skills resolves once merged

About TensorFeed

TensorFeed.ai is an AI news aggregator + machine-readable data hub for AI agents. Public source at https://github.com/RipperMercs/tensorfeed. Endpoint catalog at https://tensorfeed.ai/api/meta. Agent-friendly entry doc at https://tensorfeed.ai/llms.txt. Also available as a connector in anthropics/life-sciences#41 for the Life Sciences vertical marketplace.

Adds the TensorFeed skill teaching Claude when and how to use
TensorFeed.ai's hosted MCP server (https://tensorfeed.ai/api/mcp)
for live, machine-readable AI ecosystem data.

The skill covers 17 free MCP tools across:
  - News + AI ecosystem (3 tools)   live news, status, models
  - Security advisories  (5 tools)   CVE, KEV, EPSS, OSV (x2)
  - Finance + filings    (3 tools)   SEC EDGAR search/submissions/tickers
  - FDA regulatory       (5 tools)   FAERS, SPL, recalls (x2), MAUDE
  - Energy + macro       (1 tool)    EIA series

Trigger guidance: AI news / model pricing comparison / AI service
status / security CVE lookup / FDA recall lookup / SEC company
filing / energy or macro indicator. Skip: pure code generation
without data dependency, the user wants their own private data,
non-AI-domain questions.

License posture: most data is US Government public domain or CC0
(openFDA, BLS, FRED, MITRE, CISA, EIA, SEC); commercial
redistribution permitted; attribution preserved on every response.

The skill registers as a standalone plugin entry mirroring the
claude-api pattern. SKILL.md is 6.2KB; the body covers connection,
trigger conditions, the tool catalog grouped by domain, output
discipline (ok/source/fetched_at/attribution), and notes the optional
premium REST tier without recommending it unless explicitly asked.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants