ETC
ETC OS · Apollo
Research Tool
← Opportunities
Audit · MCP Servers · Current

MCP inventory & gap analysis

A reconciliation of every MCP server we discussed across the design conversations against what is live today. The Public Data Tier (Wave 1) is now integrated and registered in both CLIs — 14 MCPs are operational. This page tells you what is live, what is conditional (Wave 2), what stays deferred to Phase 2, and what is out of scope.

✓ Integrated & live
14
8 core/custom + 6 Public Data Tier
Wave 2 — conditional
3
Paid scrapers — activate per avenue priority
Phase 2 (deferred)
3
Internal systems + regulatory wrappers
Out of scope (for now)
~
Overlap with live tools — revisit if needed
✓ Wave 1 integrated — DEC-006

The six Public Data Tier MCPs are now live and registered in both MCP registries (.claude/settings.json + .codex/config.toml). CLAUDE.md carries the canonical Tool Dispatch table that tells the agent which MCP to reach for, and the avenue-researcher skill encodes the specialized-MCP cascade. All six are free-tier.

✓ tavily ✓ paper-search-mcp ✓ patent-connector ✓ edgartools ✓ octagon-sec ✓ gov-ca-mcp
Legend: ✓ Shipped 📋 Phase 2 (planned) 💭 Mentioned only ⊝ Out of scope Effort: S ½–1 day · M 2–5 days · L 1–2 weeks

✓ Integrated & live (14)

All built/registered and tested, live in both .claude/settings.json and .codex/config.toml. The first eight shipped in v0; the six Public Data Tier MCPs were integrated under Wave 1 (DEC-006).

📋 Deferred to Phase 2 (planned)

These MCPs were explicitly scoped out of v0 in §6 of the IMPLEMENTATION_PLAN.md ("Phase 2 — Out of v0 scope"). They cover internal-systems integration and regulatory data sources.

💭 Not integrated — paid or overlapping (4)

These came up during earlier research turns as "candidate MCPs in the ecosystem." Tavily, EdgarTools, and Octagon have since been integrated (see above), and patents are now covered by patent-connector. What remains is either a paid Apify scraper (Wave 2 — activate per avenue priority) or overlaps with a tool already live.

🤝 What's resolved · what's still open

The Public Data Tier questions (Q3, Q4) are resolved — those MCPs are integrated and live. Two questions remain genuinely open, both deferred to Phase 2 because they depend on ETC-internal decisions. I'll write a new PROPOSAL in coordination/PROPOSALS.md for each accepted item.

  1. Q1: Internal systems — etc-sharepoint & etc-crm

    The Avenue Framework's highest-converting avenues (Direct VOC, Co-Development, Customer Support & Field Failure, Channel Partner Signals) all draw from ETC's internal systems. Today there's no MCP for them, so those avenues fall back to manual data.

    Which systems does ETC actually use? (SharePoint version? CRM = HubSpot / Salesforce / Dynamics / other? Where do support tickets live?)

  2. Q2: Regulatory wrappers — which agencies first?

    There are 8 agencies on the list (CFIA, ISED, FCC, NRCan, ISC, PSPC, AER, ECCC). Each is a small scraper MCP. Building all 8 is medium effort, building one is small.

    Which 1–2 agencies map to ETC's active strategic priorities (mining vs. agriculture vs. defence vs. infrastructure)?

  3. ✓ Q3: Public-data enrichment — RESOLVED

    EdgarTools (SEC filings) and Octagon (financials) are integrated and live for the M&A / Strategic Acquisition avenue; patent-connector covers the Patent Landscape avenue. Only the paid Apify PitchBook/Crunchbase scrapers remain conditional (Wave 2), to be activated when an M&A avenue is explicitly prioritized.

  4. ✓ Q4: Tavily fallback — RESOLVED

    Wired. tavily is registered in both CLIs as the search hedge/fallback when exa rate-limits or misses.

  5. Q5: Anything missing that I haven't captured?

    If during your read you spot an MCP we discussed that I didn't list — flag it. I'll add it to the audit and resolve its status.

How to respond: just answer in plain language. For each question, tell me which way you want to go and I'll write the formal PROPOSAL in coordination/PROPOSALS.md and update the roadmap to reflect the accepted scope.