Seer MCP proxy refinement — 2026-06-16
Purpose: refine which released Seer skills (marketplace-listed) are plausibly proxied by MCP-call logs. Released skills counted: 66 (preview/experimental excluded).
MCP-observable expected
33/66 — 50.0%
MCP-optional / undercount risk
17/66 — 25.8%
Not a good MCP proxy
16/66 — 24.2%
Skill classification and likely proxy signals
For each released skill, bucket and a concise reason explaining why an MCP call is likely (or why it may be optional / excluded). Signals referenced: BigQuery / Seer Signals queries, DataForSEO API calls, Wrike updates, Looker reports, Drive/Guru artifacts, and other Seer data-system interactions.
| Skill | Bucket | Why observable / proxy signal | Path |
|---|---|---|---|
| core-dependencies/core-quality-standards | MCP-observable expected | Runs quality checks against stored datasets; emits Seer Signals / BigQuery queries and validation logs visible in MCP telemetry. | plugins/core-dependencies/skills/core-quality-standards |
| core-dependencies/core-writing-standards | Not a good MCP proxy | Guidance and authoring standards only; no backend data calls expected. | plugins/core-dependencies/skills/core-writing-standards |
| core-dependencies/core-presenting | MCP-optional / undercount risk | Presentation generation may pull Drive templates or Looker snippets (MCP) but can run from user-provided artifacts; MCP use is situational. | plugins/core-dependencies/skills/core-presenting |
| core-dependencies/core-marketing-value-types | MCP-optional / undercount risk | May enrich with market data (BigQuery/Signals) for benchmarks; optional when users supply context. | plugins/core-dependencies/skills/core-marketing-value-types |
| core-dependencies/core-qa-check | MCP-observable expected | Automated checks query datasets and scoring systems (Seer Signals / BigQuery) — strong MCP-call footprint. | plugins/core-dependencies/skills/core-qa-check |
| core-dependencies/core-seer-branding | Not a good MCP proxy | Branding guidance and static assets; no data-system calls expected. | plugins/core-dependencies/skills/core-seer-branding |
| core-dependencies/seer-skills-help | Not a good MCP proxy | Help and documentation surfaces only; not data-driven. | plugins/core-dependencies/skills/seer-skills-help |
| core-dependencies/core-chat-handoff | Not a good MCP proxy | Workflow handoff, metadata-only; MCP calls not required for normal use. | plugins/core-dependencies/skills/core-chat-handoff |
| core-dependencies/data-client-conversations | MCP-observable expected | Loads client conversation history from data stores (Seer Signals / BigQuery/Drive) — likely MCP activity when used. | plugins/core-dependencies/skills/data-client-conversations |
| seo/seo-content-audit | MCP-observable expected | Content audits call SERP and DataForSEO feeds and Seer Signals for traffic data — BigQuery/DataForSEO calls appear in MCP logs. | plugins/divisions/seo/skills/seo-content-audit |
| seo/seo-content-gap | MCP-observable expected | Fetches SERP results and keyword datasets (DataForSEO / BigQuery) for gap analysis — generates MCP calls. | plugins/divisions/seo/skills/seo-content-gap |
| seo/seo-content-outline | MCP-observable expected | Produces outlines using keyword and topic signals pulled from DataForSEO / Signals — observable MCP enrichment queries. | plugins/divisions/seo/skills/seo-content-outline |
| seo/seo-internal-linking | Not a good MCP proxy | Authoring pattern guidance; internal link suggestions may be local and not trigger MCP calls. | plugins/divisions/seo/skills/seo-internal-linking |
| seo/seo-methods | MCP-observable expected | Method runbooks that include data pulls (SERP snapshots / historical traffic) — typical run includes MCP queries. | plugins/divisions/seo/skills/seo-methods |
| seo/seo-quick-wins | MCP-observable expected | Quick-win analysis often queries site analytics (Seer Signals / BigQuery) to score opportunity — shows in MCP logs. | plugins/divisions/seo/skills/seo-quick-wins |
| seo/seo-search-landscape | MCP-observable expected | Landscape reports pull keyword and SERP histories (DataForSEO/BigQuery) — clear MCP-call footprint. | plugins/divisions/seo/skills/seo-search-landscape |
| seo/seo-serp-trends-base | MCP-observable expected | SERP trend calculations rely on historical datasets (BigQuery / Seer Signals) — observable queries expected. | plugins/divisions/seo/skills/seo-serp-trends-base |
| seo/seo-systematic-investigation | MCP-optional / undercount risk | May hit external APIs for deep investigation but can run off uploaded data; MCP calls optional depending on workflow. | plugins/divisions/seo/skills/seo-systematic-investigation |
| seo/seo-traffic-decline-analysis | MCP-observable expected | Analyses use analytics backends (Looker/BigQuery/Seer Signals) to compare windows — generates MCP telemetry. | plugins/divisions/seo/skills/seo-traffic-decline-analysis |
| seo/tseo-methods | Not a good MCP proxy | Technical SEO methods documentation; mostly guidance without data pulls. | plugins/divisions/seo/skills/tseo-methods |
| pdm/pdm-methods | MCP-observable expected | Product methods that run experiments against metrics in BigQuery/Looker — routine MCP queries expected. | plugins/divisions/pdm/skills/pdm-methods |
| pdm/pdm-weekly-optimization | MCP-observable expected | Weekly optimizations pull product analytics (BigQuery) and produce update artifacts — generates MCP activity. | plugins/divisions/pdm/skills/pdm-weekly-optimization |
| pdm/pdm-testing | MCP-observable expected | Testing harness reads test results in data stores and posts summaries (BigQuery / Seer Signals) — observable in MCP logs. | plugins/divisions/pdm/skills/pdm-testing |
| pdm/pdm-sqa | MCP-observable expected | SQA checks integrate with data and ticket systems (Looker / Wrike) producing MCP calls for traces and evidence. | plugins/divisions/pdm/skills/pdm-sqa |
| pdm/pdm-competitive-analysis | MCP-observable expected | Competitive pulls use external datasets and DataForSEO/BigQuery to build reports — shows in MCP telemetry. | plugins/divisions/pdm/skills/pdm-competitive-analysis |
| pdm/pdm-keyword-analysis | MCP-observable expected | Keyword analysis queries DataForSEO / Signals datasets; calls are visible in MCP logs. | plugins/divisions/pdm/skills/pdm-keyword-analysis |
| pdm/pdm-report | MCP-observable expected | Reports compile Looker/BigQuery extracts and Drive artifacts — typical run touches MCP data systems. | plugins/divisions/pdm/skills/pdm-report |
| analytics/ana-event-tracker | MCP-observable expected | Instrumented event tracking writes to BigQuery / Seer Signals; skill runtime will call MCP endpoints for event data. | plugins/divisions/analytics/skills/ana-event-tracker |
| analytics/ana-first-party-data | MCP-optional / undercount risk | May query first-party stores (BigQuery) when available; if user supplies CSV, MCP calls may not occur — optional. | plugins/divisions/analytics/skills/ana-first-party-data |
| analytics/ana-funnel-analysis | MCP-observable expected | Funnel analysis pulls conversion events from Seer Signals / BigQuery — generates MCP queries. | plugins/divisions/analytics/skills/ana-funnel-analysis |
| analytics/ana-kpi-development | MCP-optional / undercount risk | KPI design may consult Looker/BigQuery but can be done offline; MCP calls depend on data needs. | plugins/divisions/analytics/skills/ana-kpi-development |
| analytics/ana-measurement-strategy | MCP-observable expected | Strategy work references measurement data and Looker extracts — typical runs touch MCP data stores. | plugins/divisions/analytics/skills/ana-measurement-strategy |
| analytics/ana-methods | MCP-observable expected | Method templates often execute sample queries against analytics datasets (BigQuery) — observable in MCP logs. | plugins/divisions/analytics/skills/ana-methods |
| analytics/ana-reporting-buddy | MCP-observable expected | Reporting buddy auto-generates Looker/BigQuery extracts and Drive reports — MCP calls for data pulls and Drive writes are visible. | plugins/divisions/analytics/skills/ana-reporting-buddy |
| creative/cre-audience-analysis | MCP-optional / undercount risk | Audience analysis can call Signals/BigQuery for segments but often runs from provided research — MCP use varies. | plugins/divisions/creative/skills/cre-audience-analysis |
| creative/cre-brand-content | MCP-optional / undercount risk | Brand content generation may enrich from Drive/Guru or external data; MCP calls optional when enrichment enabled. | plugins/divisions/creative/skills/cre-brand-content |
| creative/cre-content-research-writer | MCP-optional / undercount risk | Research writer may fetch citations or Drive notes; can run offline — MCP calls not guaranteed. | plugins/divisions/creative/skills/cre-content-research-writer |
| creative/cre-cro-audit | MCP-observable expected | CRO audits query analytics datasets (BigQuery/Seer Signals) and session data — expected MCP telemetry. | plugins/divisions/creative/skills/cre-cro-audit |
| creative/cre-customer-journey | MCP-optional / undercount risk | Journey mapping may use Looker/Drive for artifacts; MCP use depends on whether system data is consulted. | plugins/divisions/creative/skills/cre-customer-journey |
| creative/cre-methods | Not a good MCP proxy | Method docs and playbooks; not data-driven by default. | plugins/divisions/creative/skills/cre-methods |
| creative/cre-playbook | MCP-optional / undercount risk | Playbook may reference Drive templates or query examples; actual MCP calls depend on usage pattern. | plugins/divisions/creative/skills/cre-playbook |
| creative/cre-sea-of-sameness | Not a good MCP proxy | Conceptual critique content; not expected to call data systems. | plugins/divisions/creative/skills/cre-sea-of-sameness |
| operations/ops-meeting-insights-analyzer | MCP-optional / undercount risk | May call transcript stores or Drive notes for enrichment; if provided, MCP calls optional. | plugins/divisions/operations/skills/ops-meeting-insights-analyzer |
| operations/ops-methods | MCP-observable expected | Ops methods that run checks against system data and schedules (Wrike/BigQuery) — MCP calls expected. | plugins/divisions/operations/skills/ops-methods |
| operations/ops-seerversary-card | Not a good MCP proxy | Generative card creation and templates; no backend calls required. | plugins/divisions/operations/skills/ops-seerversary-card |
| operations/ops-wrike-time-tracking | MCP-observable expected | Integrates with Wrike time APIs and updates tickets — Wrike/Seer MCP activity visible. | plugins/divisions/operations/skills/ops-wrike-time-tracking |
| innovation/inn-discovery-intake | MCP-optional / undercount risk | Intake may pull context from Drive/Guru or Signals; optional MCP involvement depending on setup. | plugins/divisions/innovation/skills/inn-discovery-intake |
| innovation/inn-feedback-sweep | MCP-optional / undercount risk | Feedback sweeps often query ticket systems or Drive; MCP calls happen when integrated, otherwise optional. | plugins/divisions/innovation/skills/inn-feedback-sweep |
| innovation/inn-seer-agent-skill-developer | Not a good MCP proxy | Developer tooling and docs; not a data-call-producing skill in normal use. | plugins/divisions/innovation/skills/inn-seer-agent-skill-developer |
| innovation/inn-slack-research | Not a good MCP proxy | Slack-focused research patterns and exports; typically not routed through MCP data pipelines. | plugins/divisions/innovation/skills/inn-slack-research |
| innovation/inn-session-wrapup | Not a good MCP proxy | Session summary generator using provided notes; not expected to hit MCP data systems routinely. | plugins/divisions/innovation/skills/inn-session-wrapup |
| innovation/inn-opencode-mechanic | Not a good MCP proxy | Mechanic/tooling skill for local ops; not data-driven for normal runs. | plugins/divisions/innovation/skills/inn-opencode-mechanic |
| data-fieldguides/data-signals-data | MCP-observable expected | Directly interfaces with Signals and BigQuery to surface schemas and queries — strong MCP-call signature. | plugins/data-fieldguides/skills/data-signals-data |
| data-fieldguides/data-signals-lookml | Not a good MCP proxy | LookML guidance and examples; mostly docs and templates, not runtime data pulls. | plugins/data-fieldguides/skills/data-signals-lookml |
| data-fieldguides/data-signals-source-matrix | MCP-optional / undercount risk | Source matrix may reference external sources; actual MCP calls depend on whether the skill pulls live data. | plugins/data-fieldguides/skills/data-signals-source-matrix |
| bd/bd-lead-research-assistant | MCP-optional / undercount risk | Lead research can call external data APIs but also accepts pasted lead lists; MCP use varies. | plugins/divisions/bd/skills/bd-lead-research-assistant |
| cs/cs-client-analysis | MCP-observable expected | Client analysis pulls CRM and Signals data (BigQuery/Drive) to assemble reports — generates MCP queries. | plugins/divisions/cs/skills/cs-client-analysis |
| cs/cs-qbr-strategic-meeting | MCP-observable expected | QBR generation reads account metrics (Looker/BigQuery) and Drive slides; MCP activity expected for data pulls and Drive writes. | plugins/divisions/cs/skills/cs-qbr-strategic-meeting |
| cs/cs-account-planning | MCP-observable expected | Account planning integrates account data (BigQuery/CRM) to produce plans — typical run triggers MCP calls. | plugins/divisions/cs/skills/cs-account-planning |
| cs/cs-methods | Not a good MCP proxy | Methodology docs for CS playbooks; not data-call-heavy. | plugins/divisions/cs/skills/cs-methods |
| seer-xdiv/xdiv-client-context-loader | MCP-observable expected | Loads client context from Seer XDiv stores / BigQuery — designed to call MCP-backed context loaders. | plugins/seer-xdiv/skills/xdiv-client-context-loader |
| seer-xdiv/xdiv-client-intelligence-research | MCP-optional / undercount risk | Research tasks may fetch external sources (Reddit, BigQuery) but can run with supplied files — MCP optional. | plugins/seer-xdiv/skills/xdiv-client-intelligence-research |
| seer-xdiv/xdiv-reddit-brand-monitor | Not a good MCP proxy | Monitoring that uses Reddit scraping or third-party APIs often outside MCP; not a reliable MCP proxy. | plugins/seer-xdiv/skills/xdiv-reddit-brand-monitor |
| seer-xdiv/xdiv-universal-reviewer | MCP-observable expected | Reviewer aggregates signals across datasets (BigQuery/Drive/Guru) — typically triggers MCP calls for lookups. | plugins/seer-xdiv/skills/xdiv-universal-reviewer |
| px/px-performance-review | MCP-observable expected | Performance reviews pull people metrics and dashboards (Looker/BigQuery); runs produce MCP telemetry. | plugins/divisions/px/skills/px-performance-review |
| px/px-people-feedback | MCP-optional / undercount risk | Feedback collection can post/read Drive forms or CRM but also accept direct input; MCP involvement varies. | plugins/divisions/px/skills/px-people-feedback |
Methodology: this refinement examined released marketplace skills only (from .claude-plugin/marketplace.json); preview and experimental skills excluded. "MCP-observable expected" means a normal primary run usually causes Seer MCP / data-system calls (BigQuery, DataForSEO, Looker, Wrike, Drive, Guru). "MCP-optional" means MCP calls may happen depending on workflow or whether the user supplies data. "Not a good MCP proxy" are guidance/method skills that do not reliably invoke MCP calls. Use MCP-call logs as a proxy for data-connected workflows (recommended: 33/66 = 50.0%) but add explicit telemetry to measure total adoption.