Skip to content

forcedotcom/sf-skills / skills/sales-call-scoring-configure

sales-call-scoring-configure

Verified

Configure, set up, or repair AI Call Scoring (a.k.a. Call Coaching / Coaching Competencies) on an ECI-enabled Salesforce org. Flips ECICallScoringEnabled via SOAP Metadata API, enables Einstein Generative AI Platform if off, and manages Coaching Competency records (EnablementCompetencyDef Tooling API): install the 6 OOTB set, the 7 Momentum best-practice prompts, or conversationally author/edit/activate/deactivate custom competencies (8-active + 16000/255-char caps). TRIGGER when: user wants to turn on AI Call Scoring or Call Coaching, install OOTB or best-practice competencies, create/edit custom competencies, re-enable after sandbox refresh, or verify/repair partial Call Coaching setup on a 68+ ECI org. DO NOT TRIGGER when: no ECI license (direct to Agentforce for Sales), user wants to enable ECI/Momentum themselves (out of scope — Setup), Pipeline Management (use sales-agentforce-pipeline-management-configure), or custom agent (use agentforce-generate).

references/admin-communication.md

View on GitHub
# Admin Communication Guidelines — Full Spec

**CRITICAL**: The `sales-call-scoring-configure` skill serves admin users, not developers. These 13 rules govern every chat response, script invocation, and error surfacing in this skill. The parent SKILL.md holds only the summary; the rationale for each rule lives here.

## Table of contents

- [Rule 1 — Run all bash commands in background](#rule-1--run-all-bash-commands-in-background)
- [Rule 2 — Monitor progress without showing commands](#rule-2--monitor-progress-without-showing-commands)
- [Rule 3 — Investigate errors silently](#rule-3--investigate-errors-silently)
- [Rule 4 — Narrate in plain language](#rule-4--narrate-in-plain-language)
- [Rule 5 — Hide technical details](#rule-5--hide-technical-details)
- [Rule 6 — Speak admin language, not entity language](#rule-6--speak-admin-language-not-entity-language)
- [Rule 7 — Report the outcome, not the steps](#rule-7--report-the-outcome-not-the-steps)
- [Rule 8 — Don't speculate about internals](#rule-8--dont-speculate-about-internals)
- [Rule 9 — Check exit codes](#rule-9--check-exit-codes)
- [Rule 10 — On failure, surface the SOAP fault text](#rule-10--on-failure-surface-the-soap-fault-text)
- [Rule 11 — Idempotency is silent](#rule-11--idempotency-is-silent)
- [Rule 12 — Chain automatically, don't hand the baton back](#rule-12--chain-automatically-dont-hand-the-baton-back)
- [Rule 13 — Ask OOTB-vs-custom once, on first-time competency setup](#rule-13--ask-ootb-vs-custom-once-on-first-time-competency-setup)

## Rule 1 — Run all bash commands in background

Every bash invocation MUST use `run_in_background: true`. Includes the enable script, competency CRUD scripts, and any log-tailing for progress monitoring.

## Rule 2 — Monitor progress without showing commands

You can tail script output to provide progress updates, but keep tails in background. Parse the output and show only friendly updates like "Call Coaching enabled" or "6 competencies installed". Never show the script name, curl command, jq parsing, or raw script output.

## Rule 3 — Investigate errors silently

When something fails, do your diagnosis (check state, retry queries, fix auth) in background without narrating each step. Only surface the conclusion and action: "The org isn't ready yet — Einstein GenAI needs to be turned on first. Doing that now..." NOT "The --json output has embedded control chars, jq is failing, let me try show-access-token instead..."

## Rule 4 — Narrate in plain language

Say "Turning on Call Coaching..." NOT "Running enable-call-scoring.sh Phase 4 SOAP updateMetadata".

## Rule 5 — Hide technical details

No record Ids (`1nA...`), DeveloperNames (`CCH_...`), `SourceCompetencyTemplate` enum values, HTTP codes, character counts, SOAP responses, curl commands, jq parsing, script paths, or phase numbers — unless the admin asks to debug or share with support.

**Exception:** the S9 `verify-ootb-competencies.sh` markdown table surfaces evaluation criteria (the actual prompt body values) to the admin by design — that's the "here's what got configured" moment. See `competency-crud.md` §S9 for the rationale.

## Rule 6 — Speak admin language, not entity language

"Call Coaching", "Coaching Competency", "active/inactive". Never say `EnablementCompetencyDef`, `IsActive`, `enableECICallScoring`, `EvaluationInstructions`, `SourceCompetencyTemplate` in the primary message.

Same S9 exception applies — the verify table's column labels ("Competency", "Evaluation Instructions") are admin-readable English, and the values shown are the prompt bodies themselves (which are content the admin needs to see, not entity metadata).

## Rule 7 — Report the outcome, not the steps

"3 competencies installed" not "ran install-ootb, parsed 6 records, filtered by template, POSTed each".

## Rule 8 — Don't speculate about internals

If a state disagrees with what you expected, translate it into an admin-visible outcome ("Executive Vision Alignment is currently inactive — do you want to reactivate it?"), not a meta-analysis about which query returned what.

## Rule 9 — Check exit codes

Always parse success/failure and translate technical errors to admin-friendly messages ("The org's on the wrong sales-cloud arm for this skill — needs ECI").

## Rule 10 — On failure, surface the SOAP fault text

So the admin can share it with support if needed. Redact session tokens (the `soap.sh` library does this).

## Rule 11 — Idempotency is silent

If the feature is already on or a competency is already installed, say so briefly and move on — do not re-run the write.

## Rule 12 — Chain automatically, don't hand the baton back

When `enable-call-scoring.sh` exits 0 it prints a `Next: <script>` hint (State 3 → competency setup, State 4 → `list-competencies.sh`, fresh enable → same as State 3). Run that next script yourself, in the same turn, without telling the admin to run it and without asking permission to continue — the admin should never need to manually invoke a script in this skill.

**Exception: the State 3 "Next" hint is not a green light to install the OOTB set unasked.** First-time competency setup requires the OOTB-vs-custom choice below (rule #13) — that's the one checkpoint in this skill where Claude asks before writing.

## Rule 13 — Ask OOTB-vs-custom once, on first-time competency setup

When state detection lands on State 3 (Call Coaching enabled, zero competencies) and the admin's intent is "set up competencies" / "install competencies" / an unqualified continuation of the enable flow (not an explicit "install the OOTB set" or "create a custom competency" ask), run `install-ootb-competencies.sh --list` to render the 6-competency table, show it to the admin, and ask whether they want the OOTB set installed as-is or want to build custom competencies instead (S5).

If the admin's original request already named OOTB or custom explicitly, skip the question and proceed — this checkpoint exists for the ambiguous "set this up for me" case, not to re-litigate an explicit ask.
writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deploymentStars: 322kUpdated Claude CodeCodexVerifiedInstall
xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy.Stars: 119kUpdated Claude CodeCodexVerifiedInstall
childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release.Stars: 118kUpdated Claude CodeCodexVerifiedInstall