connect-signals — Claude Code command: what it does, how to install
connect-signals sets up data connections, schema alignment, and monitoring for B2B intent, product, and engagement feeds.
Apache-2.0 ★ 450 updated 2026-04-03
.claude/commands/connect-signals.mdWhat connect-signals does
This command provisions data pipelines across external and internal streams such as Bombora, 6sense, G2, or app telemetry. It executes a five-stage sequence: credential inventory, taxonomy normalization, warehouse ingestion setup with deduplication, data quality monitoring, and table exposure for downstream prioritization.
Deliverables include a connection manifest with SLA tracking, warehouse schema documentation with sample SQL queries, and webhook notification settings. The orchestrator saves every task dependency and audit record directly into .claude/plans/plan-<timestamp>.json.
Ingestion failures stop after two retries before triggering an escalation ladder from data engineer to analyst. Missing vendor API credentials or target warehouse identifiers will halt pipeline provisioning at step one.
What it helps with
- ingest multi-source intent data and product feeds into unified schemas
- define warehouse deduplication logic and retention policies
- configure automated data freshness monitors and failure webhooks
- prepare normalized signal tables for account scoring workflows
Example prompts for the agent
- /intent-signal-orchestration:connect-signals --sources "bombora,g2,product" --accounts enterprise --topics "security,automation"
- /intent-signal-orchestration:connect-signals --sources "6sense,web" --window 30d --warehouse analytics_prod
- /intent-signal-orchestration:connect-signals --sources "community,product" --accounts midmarket
How to install connect-signals
Manually: put the files in place
.claude/commands/connect-signals.mdCall it with /connect-signals. For all projects use ~/.claude/commands/
Manually: put the files in place
~/.codex/prompts/connect-signals.mdCall it with /prompts:connect-signals. Codex marks this format deprecated in favour of skills
Manually: put the files in place
.github/prompts/connect-signals.prompt.mdManually: put the files in place
.gemini/commands/connect-signals.tomlGemini stores commands as TOML: the prompt text goes into the prompt field
Questions about connect-signals
How to install connect-signals in Claude Code?
Place the command file at .claude/commands/connect-signals.md inside your project directory.
What does connect-signals do?
It orchestrates data ingestion from intent and product sources, aligns destination schemas, and sets up freshness monitors.
Where does connect-signals store its execution plan?
The run details, dependencies, and audit logs are saved as a JSON plan in .claude/plans/plan-<timestamp>.json.
Source text
connect-signals.md · Apache-2.0
--- name: connect-signals description: Sets up data connections, schemas, and monitoring for intent, product, and engagement feeds. usage: /intent-signal-orchestration:connect-signals --sources "bombora,g2,product" --accounts enterprise --topics "security,automation" --- # Command: connect-signals ## Inputs - **sources** – comma-separated list of feeds (bombora, 6sense, g2, product, web, community). - **accounts** – optional filter (enterprise, midmarket, smb) or named account list file. - **topics** – optional topics to whitelist for ingestion. - **window** – lookback period (e.g., 30d) for initial sync. - **warehouse** – destination schema or dataset identifier. ### GTM Agents Pattern & Plan Checklist > Mirrors GTM Agents orchestrator blueprint @puerto/plugins/orchestrator/README.md#112-325. - **Pattern selection**: Signal connection typically runs **pipeline** (inventory → schema → pipelines → quality → activation). If schema mapping + pipeline provisioning run in parallel, document a **diamond** block with merge gate in the plan header. - **Plan schema**: Save `.claude/plans/plan-<timestamp>.json` capturing source inventory, auth, task IDs, dependency graph (data eng, security, RevOps), error handling, and success metrics (freshness, coverage %, SLA compliance). - **Tool hooks**: Reference `docs/gtm-essentials.md` stack—Serena for schema diffs, Context7 for API docs + vendor SLAs, Sequential Thinking for rollout reviews, Playwright for any UI-based connector QA. - **Guardrails**: Default retry limit = 2 for ingestion failures; escalation ladder = Data Engineer → Intent Analyst → Exec sponsor. - **Review**: Run `docs/usage-guide.md#orchestration-best-practices-puerto-parity` before handoff to confirm dependencies + approvals logged. ## Workflow 1. **Inventory & Access** – validate API keys, service accounts, and throttling limits per source. 2. **Schema Alignment** – map identifiers, normalize fields, and define topic taxonomy. 3. **Pipelines & Storage** – configure ingestion jobs, dedupe logic, and retention policies in the warehouse. 4. **Data Quality Gates** – set freshness monitors, anomaly detection, and alert routing. 5. **Activation Hooks** – expose curated tables/feeds to `prioritize-accounts` and downstream automations. ## Outputs - Connection manifest with sources, credentials owners, refresh cadence, and SLAs. - Warehouse schema doc plus sample queries for RevOps/data teams. - Monitoring dashboard setup instructions + webhook alerts. - Plan JSON entry stored/updated in `.claude/plans` for audit trail. ## Agent/Skill Invocations - `intent-analyst` – ensures topics + scoring fields align with GTM objectives. - `automation-lead` – configures pipelines and alerting. - `signal-scoring` skill – validates data completeness for scoring weights. ## GTM Agents Safeguards - **Fallback agents**: document substitutes (e.g., automation-lead covering analyst) when owners unavailable. - **Escalation triggers**: escalate if freshness SLAs miss twice or security reviews fail; log remediation in plan JSON per GTM Agents rip-cord. - **Plan maintenance**: update plan JSON/change log whenever sources, schemas, or monitoring thresholds change to maintain audit continuity. ---
© gtmagents · Apache-2.0 · gtmagents/gtm-agents