agentleFS
link7373link7373/agentic-sales-marketing/.claude/skills/connect-tools/SKILL.md

connect-tools

Skill1 starsChanged 3 months agoMITView on GitHub

Discover, recommend, connect, and configure external sales/marketing tools (CRM, email platforms, schedulers, ad managers, analytics, creative). Run when a workflow needs a tool category that isn't connected yet, or to audit the current stack.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Register: Update knowledge/tool-registry.md: flip the catalog row's status, record tier, capabilities configured, credential location, test …

    line 39

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/connect-tools — Wire Up the Stack

Inputs: the tool category needed (or "audit"), knowledge/business-context.md (budget), knowledge/tool-registry.md (current state + the connector catalog). Agents: tools-operator leads.

The connector catalog (check here first)

This environment ships with MCP connectors and skills for real tools. Before recommending anything external, check whether the category is already covered:

CategoryFirst choiceAlso availableFallback
CRM & pipelineHubSpot MCP—knowledge/pipeline.md alone (fine at low volume)
Email marketing / flowsKlaviyo MCP—browser automation on any ESP
Outbound email (1:1)Gmail MCP—guided manual send
SEO researchAhrefs MCPBright Data seo-audit skillBright Data search
Competitive intelSimilarweb MCPBright Data competitive-intel, brand-listeningmanual research
AnalyticsAmplitude MCPSupermetrics MCP (cross-channel pulls)platform-native dashboards via browser
Creative & designCanva MCPFigma MCPdescribed briefs the operator executes
Docs / wikiNotion MCP—this repo's knowledge/
Team commsSlack MCPIntercom MCP (customer messaging)email
Project trackingLinear MCPAsana MCPknowledge/ files
Web research/scrapingBright Data skills (scrape, search, scraper-builder)—WebSearch/WebFetch

Most MCP connectors need one-time authorization by the operator (in Claude settings → connectors, or /mcp in an interactive session). Until authorized, the connector exists but can't be used — the registry tracks this as needs auth.

Steps

  1. Establish need. Which capability is required, which skill/campaign needs it, by when? Check the registry — is something already connected (or merely needs auth) that covers it? Prefer authorizing an existing connector over adding a new tool, and extending an existing tool over adding another.
  2. Recommend candidates. From the catalog first, then external options: 2–3 tools matched to budget and scale. Zero-budget operators get free tiers first. One line each: strongest point, limitation, price at their scale, and connection path (MCP / browser / manual). Recommend one.
  3. Operator picks & authorizes/signs up. MCP connectors: give the exact authorization steps and wait. New accounts needing human identity steps (email verification, phone, payment): produce guided steps and wait. A paid tier is a money decision — approval gate applies.
  4. Connect via the layered strategy. MCP first, else browser automation, else record as manual-only with guided steps for each recurring action.
  5. Configure completely. Category-appropriate setup: email → sender identity, authentication (SPF/DKIM guidance), unsubscribe footer, list import; CRM → pipeline stages mirroring knowledge/pipeline.md (revops-operator assists); analytics → property, conversion events, UTM conventions; scheduler → accounts linked, timezone.
  6. Test. Perform a harmless verifiable action (test send to operator, draft post, test event, read-only query) and confirm it landed.
  7. Register. Update knowledge/tool-registry.md: flip the catalog row's status, record tier, capabilities configured, credential location, test result, date. Log quirks in Configuration notes so fixes persist.

Before you start

  1. Required inputs: knowledge/tool-registry.md (ships seeded — if missing, restore from the catalog above).
  2. Scan knowledge/decision-log.md for tag tooling — don't re-litigate settled tool choices without new evidence.
  3. Gate check: any paid tier or subscription needs a money-gate approval; check knowledge/approvals-log.md for a covering approval first.

If things go wrong

  • Authorization can't happen this session (non-interactive, operator away) → register the row as needs auth, park the exact steps in knowledge/pending.md, and use the fallback column in the meantime.
  • Connection test fails → try the next rung of the ladder (MCP → browser → manual); log the failure mode in Configuration notes.
  • Platform demands identity verification / CAPTCHA / payment → guided manual steps for the operator; resume automatically after.

Outputs

A working, tested, registered tool connection; or a clear record of what the operator must do to finish it. Close with the report per templates/skill-report.md.

Writeback

knowledge/tool-registry.md; spend approvals to knowledge/approvals-log.md; tool choice (tag: tooling) to knowledge/decision-log.md; blocked authorizations to knowledge/pending.md.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.